The search strategy identified a total of 107 records. Following title and abstract screening, 79 studies were excluded. Consequently, 28 studies proceeded to full-text review. An additional 15 studies were excluded after full-text review, most commonly because they did not assess intestinal permeability or were based on animal models. Manual reference searching yielded eight additional articles (Supplementary Table 1) [2,7–11,15]. Therefore, 21 studies comprising 1,130 patients met the eligibility criteria and were included in this review (Supplementary Table 2). The majority of studies (n = 14; 66.7%) included only adult patients, while six studies (28.6%) included both pediatric and adult populations; none evaluated pediatric patients exclusively. The three main findings of this narrative review are summarized in Figure 1.

Fig. 1.

Key findings on intestinal permeability in the allo-HSCT setting. Allo-HSCT: Allogeneic hematopoietic stem cell transplantation;GVHD: Graft versus host disease; GI: Gastrointestinal.

The gastrointestinal barrier is a complex physical structure that separates the internal environment of the body from the intestinal lumen and exhibits selective permeability, allowing the controlled transport of compounds and microorganisms beneficial to the host while preventing the entry of harmful substances, pathogens, and food antigens [16]. It is primarily composed of three layers: (1) the mucus layer, including inner and outer sublayers that harbor the intestinal microbiota; (2) the intestinal epithelium; and (3) the lamina propria, which contains immune cells [16]. The mucus layer is formed mainly by mucins and commensal microorganisms, including bacteria, fungi, viruses, and parasites [16]. These commensal microorganisms play key roles in maintaining host homeostasis, contributing to digestion, vitamin biosynthesis, metabolite production, and immune system development (helping to distinguish self from non-self antigens) [4,16]. The second layer, the intestinal epithelium, consists of a single layer of specialized cells, including enterocytes, goblet cells, Paneth cells, enteroendocrine cells, and microfold cells [16]. Finally, the third layer, the lamina propria, comprises components of the enteric immune system, including macrophages and dendritic cells.

Microorganisms, metabolites, antigens, and dietary particles present in the intestinal lumen may traverse the gastrointestinal barrier through multiple routes, a process that underlies the concept of intestinal permeability [16,17]. This concept is multifaceted and encompasses alterations in both transcellular pathways, mediated by membrane transporters and endocytic mechanisms, and paracellular routes, which are regulated by tight junctions [16,17]. Tight junctions are protein complexes that regulate paracellular permeability by joining adjacent epithelial cells [16,17]. These tight junction complexes are composed of transmembrane proteins, such as occludin and claudins, and peripheral membrane adaptor proteins, including members of the zonula occludens family, which anchor the junctional complex to the actin cytoskeleton [16,17]. Classically, the paracellular route has been the pathway most closely associated with disease development [17]. However, accumulating evidence in recent years indicates that the transcellular route also plays a critical role in pathogenesis [17]. In the allo-HSCT setting, both transcellular and paracellular routes may be altered, and specific methodological approaches are available to assess each of these pathways [17].

In the allo-HSCT setting, numerous studies have investigated methodologies to quantify intestinal permeability. Overall, approaches to assess intestinal permeability can be broadly categorized into two principal types: 1) Probe-based methods, and 2) Biomarker-based methods. The most commonly used methodologies in the allo-HSCT setting include: 1) Citrulline (eight studies with 342 patients) [7–13]; 2) Lactulose:Rhamnose Ratio (L:R - five studies with 186 patients) [3,10,11,13,15]; 3) Lactulose:Mannitol Ratio (L:M - four studies with 165 patients) [1,12,18,19]; 4) Intestinal fatty acid–binding protein (I-FABP: four studies with 355 patients) [7,8,20,21]; 5) Lipopolysaccharide-binding protein (LBP: four studies with 302 patients) [2,3,13,20]; 6) Chromium-51–labeled ethylenediaminetetraacetic acid absorption (51Cr-EDTA: three studies with 80 patients) [22–24]; 7) Xylose:3-O-methyl-D-glucose (X:O Ratio: three studies with 74 patients) [11,13,15]; 8) Rhamnose:3-O-methyl-D-glucose (R:O Ratio: three studies with 74 patients) [11,13,15]; and 9) Regenerating islet-derived protein 3 alpha (REG3α: two studies with 175 patients). Other methodologies used in one study each included suppression of tumorigenicity (ST2) [3], occludin [3], claudin1 [3], intestinal alkaline phosphatase (IAP) [3], zonulin [1], and calprotectin [1].

Each of these methodologies has specific advantages and limitations (Supplementary Table 3). Overall, probe-based approaches rely on measuring the urinary excretion of orally administered compounds that are poorly absorbed within the gastrointestinal tract [25,26]. These methodologies include: 1) L:M Ratio, 2) L:R Ratio, and 3) the 51-CrEDTA absorption test. Although these are well-established methods that provide a direct functional assessment of intestinal permeability, they typically require active patient participation and may be poorly tolerated in individuals with extensive gastrointestinal tract injury (a common feature in the allo-HSCT setting) [1,3,12,18,27]. Furthermore, when probe-based methodologies are expressed as ratios, false-negative results may occur if absorption of both the paracellular and transcellular probes is increased proportionally, which may give the impression that intestinal permeability is normal [28]. Finally, the permeability route (paracellular vs. transcellular) of many of these probes remains unclear, thereby limiting a comprehensive understanding of the underlying mechanisms of intestinal permeability [17].

In contrast, biomarker-based methodologies represent surrogate measures of intestinal permeability and can be assessed using serum or stool samples. These approaches include citrulline, zonulin, I-FABP, LBP, calprotectin, β-defensin-2, sCD14, myosin light chain kinase 210 isoform (MLCK210), REG3α, ST2, IAP, and claudin-3. A major advantage of biomarker-based methods is their feasibility in clinical practice, as these tests are easy to perform and do not require active patient participation [26,29]. However, further validation and standardization are needed, given the variability in sensitivity and specificity reported across different clinical settings [30,31].

This heterogeneity in intestinal permeability methodologies in the allo-HSCT setting becomes problematic for several reasons, especially considering that none of these studies have properly validated these methodologies against the gold standard methodology (Ussing chambers) [17]. First, these assays assess distinct components of the intestinal barrier and therefore represent different biological concepts [1,8]. For instance, citrulline is a surrogate marker of enterocyte mass, zonulin is a tight junction modulator, REG3α is a surrogate of Paneth cell injury, and I-FABP and I-BABP are surrogates of enterocyte turnover [1,8]. Thus, although they are broadly categorized as intestinal permeability markers, each has distinct regulatory mechanisms and temporal dynamics over the course of allo-HSCT. Second, the gastrointestinal tract is a long and anatomically heterogeneous system, with different barrier characteristics along its length (e.g., small intestine vs. colon). Accordingly, some assays are more specific to certain regions: while I-FABP is expressed in enterocytes of both the small and large intestine, I-BABP is exclusively expressed in the jejunum and ileum [8]. Finally, distinct mechanisms may underlie the increased intestinal permeability at different time points over the course of allo-HSCT. In a study evaluating intestinal permeability in mice undergoing allo-HSCT, different mechanisms were identified throughout the transplant course [32]. Whereas increased intestinal permeability in the early post-transplant period was primarily related to gross epithelial damage, the increase observed during graft-versus-host disease (GvHD) propagation was driven by myosin light chain kinase 210 isoform (MLCK210)-dependent alterations in tight junction permeability [32]. This is further supported by several studies demonstrating conflicting correlations among some of these intestinal permeability surrogates (Supplementary Table 4). Taken together, these findings highlight the need for time-specific selection and validation of intestinal permeability assessment methodologies throughout the allo-HSCT course. Understanding how intestinal permeability evolves over time during allo-HSCT is essential to optimize methodology selection and to interpret results within an appropriate biological and clinical context. In the following section, the current evidence demonstrating intestinal permeability dynamics over the course of allo-HSCT will be further explored.

The dynamics of intestinal permeability were evaluated in 18 studies including 834 patients using a total of 21 different methodologies (Table 1). Probe-based methodologies were used in 12 studies involving 425 patients [1,3,10–13,15,18,19,22–24]. Furthermore biomarker-based methodologies were used in 13 studies, but evaluated in a total of 657 patients [1–3,7–13,20,32]. Across methodologies, studies consistently point toward increased intestinal permeability in the allo-HSCT setting (Supplementary Table 5).

Overall, studies using probe-based methodologies indicate that both transcellular and paracellular permeability increase over the first month following allo-HSCT. Interestingly, patients undergoing allo-HSCT already exhibited increased intestinal permeability prior to the procedure when compared with healthy controls (two studies) [1,18]. This impairment was further exacerbated over time, with a trend toward higher permeability when baseline samples were compared with those obtained at Day 0 (three studies) [11,13,15]. and Day +7 (ten studies) [1,3,10–13,15,19,22,24]. Most studies also demonstrated a persistent increase in intestinal permeability at approximately Day +14 (seven studies) [11,13,15,18,19,22,24] and Day +30 (two studies) [3,18]. Although permeability began to return toward baseline between Day +21 and Day +30, it generally remained above pre-transplant levels [12].

Biomarker-based methodologies showed heterogeneity in intestinal permeability dynamics, in some cases even yielding conflicting findings. For several indirect biomarkers, such as REG3α, citrulline, ST2, IAP, LBP, I-BABP, and I-FABP, most studies showed trends similar to those observed with probe-based methodologies, with increased permeability between Day +7 and Day +14 and values returning toward baseline around Day +30 [2,3,7,8,10–13,15,20]. In contrast, for other biomarkers, such as occludin and calprotectin, no significant dynamic changes were identified [1,3]. Nevertheless, conflicting patterns were observed for some markers. Zonulin levels were elevated at baseline compared with reference values from a control group; however, relative to baseline, they decreased significantly at Day +7 (76.45 vs. 39.95; p < 0.001), Day +14 (76.45 vs. 43.00; p = NR), and Day +21 (76.45 vs. 41.45; p < 0.001), potentially suggesting improved barrier function [1]. Similarly, β2-defensin levels also decreased over the course of allo-HSCT compared with baseline at Day +7 (17.90 vs. 7.80; p = 0.101), Day +14 (17.90 vs. 4.31; p = NR), and Day +21 (17.90 vs. 6.20; p = 0.006) [1]. These discrepancies highlight the importance of assessing multiple biomarkers and permeability pathways simultaneously in order to obtain a comprehensive perspective [33]. Therefore, future studies should validate and identify the optimal methodologies for evaluating intestinal permeability in the allo-HSCT setting. Such studies should also delineate the main factors driving increased intestinal permeability, as a better understanding of these mechanisms would enable more accurate monitoring of its dynamics throughout the allo-HSCT course.

Throughout the allo-HSCT course, several factors may act individually or synergistically to modulate intestinal permeability. Key factors consistently evaluated in allo-HSCT studies include: 1) demographic characteristics (age, sex, underlying diagnosis) [1,3,7]; 2) transplant-related variables (conditioning regimen, use of total body irradiation, time since last chemotherapy) [1,3,7,9,20,21,23,24]; 3) dietary interventions [7,10,13,19,20]; 4) intestinal microbiota [3,20]; and 5) immune response (Table 2) [9,12,13]. Significant associations were found for the following factors: 1) conditioning regimen (four studies) [9,21,23,24], 2) Hematopoietic Cell Transplantation Comorbidity Index score (HCT-CI score: one study) [1], 3) dietary interventions (three studies) [7,10,20], 4) intestinal microbiota (two studies) [3,20] and 5) immune response (three studies) [9,12,13].

It is important to note that the conflicting or nonsignificant findings related to some of these factors may reflect limitations of the individual studies (such as lack of statistical power due to small sample sizes) as well as limitations of the methodologies themselves. Each of these factors has distinct underlying mechanisms that may affect different components of the gastrointestinal barrier, ultimately resulting in increased intestinal permeability. For instance, conditioning regimens can cause extensive enterocyte injury and increase both transcellular and paracellular permeability [8]. Because both pathways may be simultaneously affected, probe-based methodologies expressed as ratios may fail to detect such changes [8,28]. Indeed, among the six studies that evaluated the impact of conditioning regimens on intestinal permeability, the two that used probe-based ratios did not identify significant associations [1,3]. In contrast, the remaining four studies, which assessed permeability using 51Cr-EDTA, citrulline, and I-FABP, found that myeloablative conditioning was associated with increased intestinal permeability [9,21,23,24]. Taken together, these data emphasize the complexity of assessing intestinal permeability in the allo-HSCT setting. Additionally, as allo-HSCT increasingly shifts toward nonmyeloablative conditioning regimens, there is a clear need to better characterize how intestinal permeability varies according to different chemotherapy protocols. Therefore, future studies should ideally use multiple complementary methodologies to capture the complex and dynamic changes that occur in the gastrointestinal barrier throughout the course of allo-HSCT.

An interesting finding of this review is the consistent association across studies between intestinal microbiota and intestinal permeability [3,20]. In the allo-HSCT setting, the intestinal microbiota has emerged as a key prognostic factor and therapeutic target, with mechanistic studies suggesting that altered intestinal permeability is one of the pathways through which the microbiota influences allo-HSCT outcomes [4,34]. This review identified that both intestinal microbiota diversity indices and microbiota composition are significantly associated with intestinal permeability [3,20]. In an allo-HSCT study including 80 patients, intestinal permeability was assessed using the L:R ratio [3]. An increased L:R ratio was significantly associated with a lower Shannon diversity index at Day +7 (p = 0.044) and Day +30 (p = 0.047). In addition, increased intestinal permeability was associated with a reduction in anaerobic organisms (p = 0.020) and with the presence of several specific bacterial species. Similarly, in another study including 117 patients, the Shannon diversity index was significantly associated with several intestinal permeability biomarkers: 1) I-FABP (baseline: r = 0.31; p = 0.043; Day +21: r = 0.45; p = 0.005), 2) LBP (Day +21: r = −0.35; p = 0.032), and 3) sCD14 (Day +21: r = −0.35; p = 0.034) [20]. Furthermore, the abundance of Blautia was significantly associated with I-FABP levels on Day +21 (r = 0.39; p = 0.017). These data highlight the close interplay between the intestinal microbiota and the gastrointestinal barrier and suggest that future therapeutic strategies targeting intestinal permeability may include modulation of the gut microbiota. Identifying potential therapeutic targets to modulate intestinal permeability is therefore of particular relevance, especially in light of its prognostic significance.

The prognostic significance of intestinal permeability in the allo-HSCT setting was evaluated in 14 studies including a total of 868 patients (Table 3) [1–3,7,11,13,15,18,20–23,35,36]. The most frequently assessed allo-HSCT–related outcomes were: 1) GVHD (ten studies with 769 patients) [1–3,7,18,20,21,23,35,36]; 2) infections (five studies with 387 patients) [1–3,13,36]; 3) overall survival (four studies with 385 patients) [3,18,20,36]; 4) non-relapse mortality (three studies with 369 patients) [7,20,36]; 5) oral mucositis (three studies with 109 patients) [1,11,13]; and 6) gastrointestinal toxicity (two studies with 35 patients) [15,22]. Additional outcomes included pulmonary complications (one study with 157 patients) [36] and transplant-associated thrombotic microangiopathy (TA-TMA; one study with 80 patients) [3].

Although GVHD was the most commonly evaluated outcome, its association with intestinal permeability remains inconsistent. Several studies using diverse permeability assessment methods (e.g., L:R ratio, L:M ratio, zonulin, I-FABP, LBP, sCD14, calprotectin, and Ussing chamber assays) and examining different clinical GVHD endpoints (incidence, gastrointestinal involvement, and severity) reported non-significant findings. Nevertheless, five (50%) of the ten available studies identified potentially relevant associations. For example, in a cohort of 45 patients, intestinal permeability was assessed at Day −7, Day +7, Day +14, and Day +21 using the L:M ratio, fecal zonulin, calprotectin, and β-defensin 2 [1]. In that study, no significant associations were observed between most markers and GVHD across the evaluated time points. However, calprotectin levels at Day +7 were significantly higher in patients who subsequently developed GVHD compared with those who did not (64.758 vs. 14.060; p = 0.040). In another study including 38 patients, those who developed severe GVHD (grade II–IV) exhibited significantly higher intestinal permeability (measured by 51Cr-EDTA) at Day +4 when compared with patients who did not develop this complication (7 vs. 11; p = 0.02) [23]. Preconditioning citrulline levels were also significantly associated with severe GVHD (grade II–IV: p < 0.01; grade III–IV: p = 0.03) in a study including 95 patients [7]. Despite these conflicting findings, the available evidence suggests that intestinal permeability may play a prognostic role in GVHD [1,2,7,21,23]. However, further clarification will require future studies with larger sample sizes and the concomitant assessment of multiple intestinal permeability markers.

Infections represent another clinically relevant endpoint, particularly in light of studies on the intestinal microbiota suggesting that bacterial translocation across the gut barrier may contribute to the development of bloodstream infections (BSI) [4,6,37]. In the allo-HSCT setting, however, among the five studies evaluating this outcome, only one reported a significant association. In that study, which included 73 patients, LBP levels were significantly higher in patients with no or mild GVHD (grade I) who developed infections compared with those who did not (p < 0.001) [2]. In contrast, within the same cohort, LBP levels were not significantly associated with infections among patients with grade II GVHD [2]. Similarly, in the remaining four studies, infections were not significantly associated with other intestinal permeability biomarkers, including the L:R ratio, fecal zonulin, L:M ratio, calprotectin, β-defensin 2, and haptoglobin genotype [1,3,13,36]. For the other clinical outcomes, conflicting and/or predominantly non-significant findings were also reported. With respect to overall survival, all available studies consistently showed no significant association with intestinal permeability [3,18,20,36]. For non-relapse mortality, among the three included studies, only one identified a significant association between REG3α levels and this outcome at the preconditioning time point (9.91 vs. 6.23; p = 0.02), Day +7 (11.51 vs. 4.20; p < 0.001), and Day +28 (17.33 vs. 10.47; p = 0.03) [7]. Predominantly non-significant associations were also reported for oral mucositis, with only one study demonstrating a significant relationship between fecal zonulin levels and this outcome at Day +14 (41.125 vs. 43.000; p = 0.048)[1]. For pulmonary complications and TA-TMA, no significant associations with intestinal permeability were identified [3,36].

Overall, these findings indicate that although intestinal permeability is frequently altered in patients undergoing allo-HSCT, its prognostic significance remains inconsistent across studies. While several investigations have reported associations between intestinal permeability and allo-HSCT–related outcomes, the available evidence is inconclusive or of limited methodological quality or both. It is possible that, for each specific clinical endpoint, a particular intestinal permeability biomarker may offer greater prognostic accuracy.

Although significant advances have been made in understanding the role of intestinal permeability in the allo-HSCT setting, much remains to be elucidated. Clear conclusions regarding the dynamics and prognostic significance of intestinal permeability during allo-HSCT remain elusive due to differences in study design, variability in biomarker assays and sampling time points, lack of adjustment for potential confounders (such as diet, medication use, and comorbidities), and the absence of comparisons with a standardized gold-standard test (Ussing chambers). Notwithstanding these limitations, this comprehensive narrative review highlights the potential role of intestinal permeability as a prognostic variable throughout the allo-HSCT course that can be monitored to guide risk stratification strategies and future therapeutic interventions. To further advance this field, future studies should perform longitudinal assessments using multiple intestinal permeability methodologies to identify the most relevant pathways and modulators of gut barrier function during allo-HSCT. Simultaneous measurement of several permeability biomarkers may better capture the heterogeneity of allo-HSCT patients. In addition, future studies should aim to systematically identify and account for all potential confounders that modulate intestinal permeability in the allo-HSCT setting, both to improve analytical accuracy and to explore them as potential therapeutic targets. Finally, mechanistic studies are needed to further elucidate the pathways through which increased intestinal permeability may influence allo-HSCT-related outcomes, including immune modulation and bacterial translocation.