The process of the search strategy and article selection for the systematic review and the meta-analysis is summarized in Fig. 1, following the PRISMA flow diagram 2020 [10]. A total of 158 articles were identified through database searches, with 25 from PubMed, 33 from CINAHL, 44 from Scopus, 13 from Cochrane and 43 from Embase. After the removal of duplicates, 111 article titles and abstracts were screened for relevance to the study’s field. Among these, 46 article titles were identified as potentially relevant. Among these, 25 article titles were excluded because only the abstract was published. The full manuscript of the remaining 21 article titles were assessed further for eligibility. Ultimately, only 10 articles met the selection criteria and were included in the systematic review. Additionally, 13 more articles were discovered by screening the citation lists of these included articles. Out of these, 8 met the inclusion criteria for this systematic review and were analysed.

One of the articles [19] included for review reported two separate trials—one involving non-pregnant women and the other involving pregnant women. We counted these as two separate trials. Therefore, a total of 19 randomized trials from 18 articles, yielded through database search and citation lists, were analyzed in this systematic review.

The 19 trials included in this systematic review were published between 2006 and 2022. Out of these, 5 of the trials reported were conducted in Italy [19,20,21,22] and 14 in Egypt [23,24,25,26,27,28,29,30,31,32,33,34,35,36,37]. The total number of participants who completed these trials was 2992 patients, with a mean Hb concentration level of ≤ 11.5 g/dL. Among them, 637 were children and adolescents, and 2355 were adults. There were 1353 participants in the LF intervention groups, and 1639 participants in the comparison/control groups. In terms of gender distribution, there were significantly more females, with 2516 (including 2266 women > 18 years and 250 children and adolescents < 18 years), as opposed to 326 males (including 89 men > 18 years and 237 children and adolescents < 18 years).

Among the 19 analyzed trials, the majority (53%—10 trials) focused on pregnant women with iron deficiency anemia (IDA) at or after 12 weeks of gestation, comprising a 2041 out of 2992 participants (939 in the intervention, 1102 in the control). This represented 68% of all participants involved in the trials [19, 21,22,23, 25, 26, 28, 29, 33,34,35]. Seven trials (37%) involved children aged over 2 [24, 27, 30,31,32, 36, 37], and one trial targeted cancer patients undergoing chemotherapy [20]. Additionally, four trials included 334 individuals with chronic health conditions (cancer, cerebral palsy, inflammatory bowel disease, obesity), all of whom had anemia (Hb < 11.5 g/dL) [20, 27, 32, 37].

Among the 1353 participants in the intervention groups, all received oral bLF. They were compared to the following groups: 1160 received ferrous sulfate [21, 22, 24,25,26,27,28,29, 33, 34, 38], 73 received ferric gluconate [20], 73 received ferrous fumarate [35], 30 received ferrous bisglycinate [30], 70 received ferric (III) hydroxide polymaltose [31, 37], 62 received iron polymaltose complex [30, 32], 33 received iron (III) hydroxide dextran complex [23], and 112 received a combination of LF and iron [24, 30, 31]. The bLF dosage ranged from 100 to 250 mg per day, and the treatment duration varied between 30 and 90 days. The study characteristics of the trials included in the systematic review are presented in Table 1.

According to the Jadad scoring scale, there was a variation in the methodological bias of the studies included in the meta-analysis, as detailed in Table 2. Studies with a Jaded score of 3 or above were considered to have low methodological bias, while those with a Jadad score of 2 or less were considered to have high methodological bias. Out of the included studies, five received 1/5 rating [29,30,31, 36, 37], six received a 2/5 rating [19, 20, 22, 25, 34, 35], six received a 3/5 rating [23, 24, 26,27,28, 32, 33], and 1 received a 4/5 rating [21]. Overall, 8 out of 18 studies were considered to have low methodological bias, while 10 out of 18 studies were considered to have high methodological bias. It is evident in the scoring process (Table 2.) that the problem originates from the lack of description of the randomisation process by most authors. The major issue, however, is the non-blinding of both patients and researchers/healthcare practitioners, which can give rise to treatment bias and may eventually influence the results. Only one study (Nappi et al., 2009) out of 18 studies described the application of an appropriate blinding process during the trial.

Out of the 19 included trials, seven were eligible for meta-analysis focusing on the effectiveness of lactoferrin versus iron supplementation in patients with a low Hb concentration profile. These trials included a total of 1397 participants: 698 supplemented with ferrous sulfate, and 699 with oral bLF. The participants were pregnant women with IDA, at a gestational age of at least 12 weeks but not more than 36 weeks.

The overall pooled SMD, using a random-effects model, revealed a statistically significant increase in Hb concentration levels in the oral bLF group in comparison with the ferrous sulfate group (SMD 0.81, 95% CI: 0.42, 1.21, p < 0.0001, I2 = 95.8%, P heterogeneity < 0.001). Based on the random-effects model, the pooled SMD, after one month of treatment was 0.59 (95% CI: 0.18, 0.99, p = 0.004; I2 = 92.1%, P heterogeneity < 0.001). For those women who received the intervention for the longest duration, the pooled SMD, was 1.04 (95% CI: 0.34, 1.74, p = 0.004; I2 = 97.1%, P heterogeneity < 0.001). The results of the meta-analysis regarding the effectiveness of bLF compared to ferrous sulfate supplementation in pregnant women with low Hb concentration are presented in Fig. 2.

Forest plot of the standardized mean differences of the Hb concentration values

Funnel plots (Fig. 3 and Fig. 4) were used for the visual inspection of asymmetry or outliers for the purpose of determining publication bias. The shape of the funnel plots indicated data asymmetry. The Egger asymmetry test revealed no evidence of publication bias (P Egger’s test P > 0.05).

Funnel plot for publication bias evaluation (1 month treatment duration)

Funnel plot for publication bias evaluation (Longest treatment duration)