Safety assessment of L-ornithine oral intake in healthy subjects: a systematic review

Abstract

L-Ornithine (L-Orn) is a nonessential amino acid but has many physiological roles. Accordingly, L-Orn has been used as a functional food or dietary supplement to ameliorate various maladies, but there is only limited information available about its safety. The safety of a chemical compound is generally assessed via non-clinical and clinical studies, but safety information derived from human studies is particularly important. Recently, systematic reviews have been used to assess the safety as well as the effectiveness and usefulness of such studies. Therefore, we conducted an assessment of the safety of L-Orn by systematically reviewing clinical studies. Specifically, we performed a comprehensive search of databases for clinical trials in which L-Orn was added to ordinary diets (i.e., orally administered) in healthy individuals. Focusing on PubMed, Cochrane Library, Ichushi-Web, and EBSCOhost, we comprehensively searched for reports on human studies on the oral ingestion of L-Orn. We identified 22 articles as subjects for this SR. Among these articles, the maximum L-Orn dose was 14,025 mg/person/day in the form of L-Orn hydrochloride and the maximum duration of administration was 156 days. The main observed adverse events were gastrointestinal disorders. Indexing these adverse events, the no observed adverse effect level was estimated to be 12,000 mg/person/day for L-Orn in the form of L-Orn hydrochloride. When we conducted an integration analysis on the risk of adverse events, the difference between those with and without L-Orn supplementation in the risk of gastrointestinal disorders was 0.00 (95% confidence interval: ±0.02, P = 1.00), so no significant effects were observed. (UMIN000033371)

Supplementary Information

The online version contains supplementary material available at 10.1007/s00726-025-03455-4.

Introduction

L-Ornithine (L-Orn) is an amino acid that is not incorporated into proteins, being present in nutritional sources as a free amino acid (Wu 2021; Wu 1998). L-Orn functions in the urea cycle in the liver, converting the toxic nitrogen metabolite ammonia to urea. It is also known as a polyamine precursor (Rodwell 1999). L-Orn is obtained by humans through certain food sources, such as brackish-water bivalves, scallops, little-neck clams (Uchisawa et al. 2004), mahi-mahi, yellowfin tuna, American red snapper (Antoine et al. 2001), and green tea (Ohtsuki et al. 1987). However, the amount provided by a typical diet is limited. While a considerable amount of L-Orn is synthesized from glutamine/glutamate and proline within the intestines, then it is converted to citrulline, which is taken up by the kidney for the synthesis of arginine (Wu 2021; Wu 1998). L- Orn has been reported to exhibit certain physiological activities; for example, its ingestion leads to the release of growth hormone (Bucci et al. 1990; Demura et al. 2010; Evain-Brion et al. 1982; Jeevanandam et al. 1996; Matsuo et al. 2015). In addition, the ingestion of L-Orn and L-arginine with exercise loading was reported to decrease weight and body fat (Elam 1988) and increase lean body mass in healthy men (Elam 1989). Because L-Orn stimulates the urea cycle (Morris 2002), it is used for pharmacological treatment to decrease blood ammonia concentrations and to reduce the symptoms of hepatic encephalopathy associated with liver cirrhosis (Butterworth et al. 2019a; Gebhardt 1997; Goh et al. 2018; Hunold 1973; Müting 1992; Vogels 1997) and non-alcoholic fatty liver disease (Butterworth et al. 2019b). L-Orn has also been reported to promote alcohol metabolism in the liver, suppress ethanol-induced hepatocellular death, and improve sleep quality after drinking (Kokubo et al. 2012, 2013; Tamai et al. 2013). In human studies, L-Orn administration has been reported to improve physical fatigue (Demura et al. 2011) and to suppress the elevation of blood ammonia during exercise (Sugino et al. 2008). As of 2018, to exploit the above-mentioned biological activities, about 170 t of L-Orn was used globally as a functional food or dietary supplement (Fuji Keizai 2019).

Regarding the safety of L-Orn, the following findings have been made. The results of an oral acute toxicity test of L-Orn in rats showed that its LD₅₀ was approximately 10 g/kg (BW) (Breglia et al. 1973). In addition, in a 13-week repeated-dose toxicity test using rats (L-Orn administration in feed at 1.25%, 2.5%, or 5.0% w/w), the no observed adverse effect level (NOAEL), including a 5-week recovery test, was 5.0% or higher [male: 3445 mg/kg (BW), female: 3986 mg/kg (BW)] (Ishida et al. 2013). Genotoxicity tests (Ames test, chromosome aberration test) were also negative for L-Orn (Ishida et al. 2013). In a reproductive toxicity test using rats, the administration of L-Orn hydrochloride in the range of 0 to 4000 mg/kg (BW) to rats for 10 days from the 5th day of pregnancy did not affect the embryo and fetus. Upon the administration of 4000 mg/kg (BW), the level of feed consumption by maternal animals showed a temporary decrease, but then recovered. No other changes of general condition, weight, or effects on major organs/tissues were observed (Kyowa Hakko Bio, unpublished data).

In general, acceptable daily intakes (ADIs)– that is, safe intake levels for humans– are derived from the no observed adverse effect level (NOAEL) determined in experimental animals through various tests that assess dose-response relationships, typically using a standard safety factor (SF) of 1/100. However, this standard approach is not applicable to determining safe intake levels for nutrients, including amino acids. For example, in the case of L-lysine, an essential amino acid, the value using the SF from a rat study is 34 mg/kg/day, which is close to the human requirement of 38 mg/kg/day, and which is significantly lower than the average intake in the USA (Blachier et al. 2021; Food and Nutrition Board (FNB) 2005). The SF of 1/100 commonly used in nutrient safety assessments is therefore inadequate for extrapolating animal data to humans, particularly for amino acid safety assessment. As such, the determination of safe levels of amino acid intake should be based on human studies (Kuramochi et al. 2023; Sakai et al. 2004; Food and Nutrition Board (FNB) 2002). A systematic review (SR) that comprehensively collects and objectively evaluates the results of clinical trials would be useful to evaluate the safety of particular levels of intake of amino acids in humans. As examples of reports on amino acid safety using SR, Hayamizu et al. conducted a human safety evaluation of L-lysine, while Kuramochi et al. did so for L-arginine (Hayamizu et al. 2019; Kuramochi et al. 2023).

With regard to L-Orn, a number of clinical trials of oral L-Orn have been conducted using a randomized controlled design as an intervention study, presumably because of its use to improve certain metabolic diseases. These reported clinical trials could therefore be used to assess the safety of L-Orn. However, to the best of our knowledge, there are no published reports of safety evaluations of L-Orn using an SR. The purpose of this SR is therefore to collect comprehensive clinical trial data and evaluate the safety of L-Orn as an add-on to a regular diet using the SR method.

Methods

We conducted an assessment of the safety of L-Orn intake targeting healthy people by SR. We set the eligibility criteria of the study (PICOS) as follows: patients (P), healthy people who took L-Orn orally; intervention (I), L-Orn; comparison (C), placebo or no treatment; outcome (O), any adverse events (AEs); and study design (S), intervention trial. We followed the Cochrane Handbook for Systematic Reviews of Interventions in conducting this meta-analysis (Higgins and Thomas 2019). The results are reported in accordance with the PRISMA 2020 statement: updated guidelines for reporting systematic reviews (Page et al. 2021). The review protocol was registered at UMIN (UMIN000033371) before beginning the study. Institutional review board approval was not required because we conducted an SR and meta-analysis, which did not directly involve human subjects.

Data source and search strategy

We searched the PubMed, Cochrane Library, EBSCOhost, and Ichushi-Web databases for studies involving L-Orn intervention in humans reported between January 1970 and March 2023. The search term was “ornithine” and results were filtered for those from a “clinical trial.” As the aim of this SR was to assess L-Orn’s safety, all intervention studies in humans were searched comprehensively without a limitation regarding the study design. No language restrictions were applied in the electronic search. Manual searches of journal articles and reference lists from relevant publications were also performed to ensure that all appropriate studies were considered for inclusion. Two investigators (HY and KH) performed the electronic search independently (Supplementary Table S1).

Inclusion/exclusion criteria

Inclusion criteria were as follows: (a) human study (on healthy humans); (b) L-Orn administered orally; (c) L-Orn used in the form of L-Orn hydrochloride (LOHC), L-Orn L-aspartate (LOLA), or L-Orn-α-ketoglutarate (LOKG); and (d) study design of intervention type. Meanwhile, exclusion criteria were as follows: (a) unknown dose of L-Orn; (b) not an L-Orn intervention study; (c) L-Orn not administered orally; (d) L-Orn salt of other acidic drugs used; (e) L-Orn derivative used; and (f) non-healthy humans. Any disagreements about which studies to include between the two investigators were resolved via discussion with another member of the reviewing team (KH, HY, YK, or SS).

Data extraction and quality assessment

Data were extracted in accordance with the study protocol (UMIN000033371). Two investigators (HY and YK) independently extracted the following data from eligible papers: (1) name of the first author, (2) year of publication, (3) study location, (4) study design, (5) numbers of participants in the L-Orn and control groups, (6) participant age, (7) L-Orn dosage per administration and per day, (8) duration of administration of L-Orn, and (9) AEs during the period of L-Orn treatment. Regarding the dosage, when L-Orn was used for intervention in the form of LOHC, LOLA, or LOKG, it was converted to net L-Orn content. When information was ambiguous or missing, we contacted the corresponding author to obtain the most accurate data available. The quality of the included studies was assessed using the Cochrane Collaboration’s tool for assessing the risk of bias (RoB) (Higgins and Green 2008), including random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other sources of bias. For each trial, the risk of bias was reported as low, unclear, or high. The criteria used for quality assessment have been described in detail elsewhere (Higgins and Green 2008). In the case of missing values or unclear points, we contacted the corresponding author of the study and added the required data whenever possible. Furthermore, the quality of clinical trials was assessed using the Jadad scale, which rates aspects of randomization, blinding, and withdrawals. This five-point quality scale includes points for randomization (randomized = 1 point; table of random numbers or computer-generated randomization = an additional 1 point), double-blinding (double-blind = 1 point; use of a placebo = an additional 1 point), and follow-up (numbers withdrawing and reasons for withdrawal in each group are stated = 1 point) as described within the reports of the randomized controlled trials. Final scores of 0–2 were considered to represent studies of low quality, whereas final scores of ≥ 3 were regarded as representing high quality (Jadad et al. 1996).

Classification of included papers

The included papers were categorized into the following three categories: Category A, no description of AEs in the paper; Category B, a statement that no adverse events occurred; and Category C, a statement that adverse events occurred during the trial (Hayamizu et al. 2019; Kuramochi et al. 2023).

Data synthesis and analysis

Safety was reviewed by comparing the frequency of AEs between the L-Orn group and the control group. The variable risk difference (RD) and 95% confidence interval (CI) were further used to calculate the pooled risk estimates. Cochran Q tests and I² statistics were used to examine the heterogeneity between studies (Higgins and Thompson 2002). The DerSimonian and Laird random-effects model was used for data synthesis of AEs because we assumed that there was heterogeneity between studies (DerSimonian and Laird 1986). Sensitivity analysis was carried out to identify the study responsible for the heterogeneity and/or to test the validity of the conclusions by omitting one study sequentially (leave-one-out test). Publication bias or small-study effect was assessed by the funnel plot method and using Egger’s test. When the P value was > 0.05 in the combined Begg’s test, it was considered that there was no publication bias (Begg and Mazumadar 1994). If there was any evidence of publication bias, the trim and fill method was used to evaluate its impact. Meta-analysis and summary of bias risk were conducted using the Cochrane Program Review Manager (RevMan) Version 5.4 (The Cochrane Collaboration 2020) and publication bias was analyzed using R4.1.1 (The R Foundation for Statistical Computing 2021) with the packages “meta” and “metafor” (Schwarzer et al. 2015; Viechtbauer et al. 2010).

Certainty of evidence

We assessed the certainty of evidence according to the Grading of Recommendations Assessment, Development and Evaluation (GRADE). This was based on four grades, namely, high quality, moderate quality, low quality, and very low quality, for the following five items: study limitations, inconsistency of results, indirectness of evidence, imprecision, and reporting bias (Guyatt et al. 2008).

Results

Of the 1227 records retrieved through electronic searches, reference searches, and manual searches, 1201 records were excluded after screening of the title and abstract, and 4 of the remaining 26 records were excluded after a full text review; thus, 22 published investigations met the inclusion criteria of the review protocol (Fig. 1). These 22 selected papers were categorized into Categories A–C. Category A included 13 articles not describing the presence or absence of AEs. Category B included 8 articles reporting that the occurrence of an adverse event was not observed. Category C included 1 article reporting that the occurrence of an adverse event was observed.

Fig. 1.

Flowchart of the selection of studies for this systematic review In Category A, AEs were not described in the article. In Category B, AEs were described, but no AEs were observed. In Category C, the occurrence of AEs was reported

Characteristics of included studies

Since the purpose of this SR was to estimate a safe daily intake, information from repeated-dose studies and information from single-dose studies are interpreted differently, so the two were considered separately. The former consisted of 10 articles and the latter consisted of 12 articles (Table 1, Supplementary Table S2).

Table 1.

Characteristics of the included studies in systematic review 【L-ornithine hydrochloride】

Study	Country	norn / ntotal	Age (year)	Dose (mg/person/day)	Dose (NET) (mg/person/day)	Dose (mg/kg/day)	Dose (NET) (mg/kg/day)	Duration (day)	Inclusion criteria	Design	Jadad score	Category
【L-ornithine hydrochloride】
Bucci 1990	USA	12/12	Male: 28.1 ± 7.0 Female: 34.3 ± 5.1	Male: 3300 8250 14025 Female: 2328 5820 9894	Male: 2582 6468 10989 Female: 1822 4563 7752	40 100 170 40 100 170	31.3 78.4 133.2 31.3 78.4 133.2	1	Bodybuilders	OL, NC	1	B B C B B C
Demura 2010	Japan	14/14	22.2 ± 1.0	7250	5684	100	78.4	1	Healthy adult males	DB, CO	2	A
Demura 2011	Japan	10/10	23.8 ± 3.9	6770	5308	100	78.4	1	Healthy adults	DB, CO	2	A
Kirisako 2012	Japan	19/39	43.5 ± 1.4	500	400	-	-	56	Healthy adult females	RCT, DB	3	B
Morita 2013	Japan	16/16	47.9 ± 2.9	3000	2351	50	39.3	90	Healthy adults	OL, NC	1	B
Komano 2013	Japan	17/17	28.7 ± 1.6	3000	2351	49	38.2	1	Healthy adult males	DB, CO	2	A
Kokubo 2013	Japan	27/27	35.5 ± 2.4	500	400	7.6	6.1	1	Healthy adults	RCT, DB, CO	3	A
Miyake 2014	Japan	26/52	43.31	500	400	-	-	56	Healthy adults	RCT, DB	3	B
Fukuda 2018	Japan	28/28	39.9	500	400	-	-	7	Healthy adults	RCT, DB, CO	5	B
Sugino 2008	Japan	17/17	40.9 ± 11.8	2000 6000	1567 4702	34.4 103.1	26.9 80.8	7 1	Healthy adults	RCT, DB, CO	3	A A
Tsai 1997	China	10/10	20	1466	1147	20	15.7	30	Healthy adult males	RCT, DB, CO	3	B
Iwasaki 1983	Japan	10/18	20–49	-	-	100	78.4	1	Healthy adults	OL	0	A
Cynober 1990	France	6/6	28 ± 1	6400	5015	87.7	68.7	1	Healthy adult males	CO, NC	1	B
Miura 2022	Japan	23/23	41.4 ± 2.1	3200 6000 9200 12000	2508 4702 7209 9403	45.8 86.0 131.8 171.9	35.9 67.4 103.3 139.0	28	Healthy adult males	RCT	2	B B B B
【L-ornithine-L-aspartate】
Kokubo 2012	Japan	31/31	Male: 34.7 ± 1.9 Female: 32.4 ± 2.3	800	398	Male: 12.2 Female:15.8	6.1 7.9	1	Healthy adults	RCT, DB, CO	3	A
Mikulski 2015	Poland	11/11	32.6 ± 1.9	12000	5970	163	81.1	1	Healthy adult males	RCT, DB, CO	3	A
Kowalski 2006	Poland	12/12	24.1 ± 2.5	3000	1493	39.7	19.7	1	Healthy adult males	RCT, CO, NC	2	A
【L-ornithine-α-ketoglutarate】
Chetlin 2000	USA	8/18	22.5 ± 1.8	10000	4762	120	57.2	42	Weight-trained males	RCT, DB	3	A
Jordan 1993	Spain	81/81	2–12	2000	952.4	-	-	156	Helthy prepubertal children of short stature	RCT	3	B
Cynober 1990	France	6/6	28 ± 1	10000	6400	137	87.7	1	Healthy adult males	CO, NC	1	B
【L-ornithine-α-ketoglutarate】
Elam 1988	USA	10/18	38.28	1000	1000	-	-	25	Healthy adult males	RCT, DB	5	A
Elam 1989	USA	11/22	37.23	1000	1000	-	-	25	Healthy adult males	RCT, DB	5	A
Zajac 2010	Poland	9/17	23.5 ± 2.1	2200	2200	25.8	25.8	21	Athletes	RCT, DB	2	A

The form of L-Orn used for administration was classified into three types: LOHC, LOLA, and LOKG, as well as other (unknown) cases in which this was not specified in the paper. Therefore, in this assessment, we performed stratified analysis for each L-Orn type. The characteristics of the included studies are summarized in Table 1. The mean age of all participants in the included studies ranged from 2 to 47.9 years. The duration of administration varied widely from 1 to 90 days, depending on the study. Therefore, in this SR, the tolerability of L-Orn was evaluated in single-dose and repeated-dose studies.

The dose of L-Orn tested in the studies ranged from 398 to 10,989 mg/person/day as the net of L-Orn, and among the included studies, 10 trials used a crossover design, 2 used a placebo-controlled parallel design, and 4 were conducted in a double-blind manner. The most common chemical form of L-Orn as an intervention sample used in the included studies was LOHC (14 articles, 20 studies). Meanwhile, LOLA and LOKG were each used in three reported studies. All studies on LOLA (Kokubo et al. 2012; Kowalski et al. 2006, Mikulski et al. 2015) were classified into Category A. Therefore, these three articles on LOLA could not be used for estimations of NOAEL and lowest observed adverse effect level (LOAEL). In three studies, the chemical form of L-Orn was unknown (Elam et al. 1988, 1989; Zajac et al. 2010). These studies were conducted on healthy subjects, but did not mention the presence or absence of adverse events, so they were classified into Category A. Therefore, these three articles could also not be used for NOAEL and LOAEL estimation.

Assessment of study quality

The quality of the included studies was heterogeneous (Fig. 2). Random sequence generation was reported in three studies. Allocation concealment was also reported in one study. The risk of potential performance bias was low in 15 studies. Among studies with available information on whether outcome assessment was blinded, the risk of detection bias was high in 6 studies. The outcome data were incomplete in 1 study. In terms of the Jadad index, 12 studies were judged as being of high quality (score ≥ 3), while 10 were of low quality.

Fig. 2.

Assessment of risk of bias for 22 selected human studies: summary of items of bias Risk of bias for all trials included in the systematic review is presented as percentages of trials with low (green), high (red), or unclear (yellow) risk of bias in each assessment item

Maximum dose, duration of administration, and sample size

Figure 3 shows an outline of the studies on healthy individuals as a bubble plot. The x- and y-axes indicate the duration of administration and the dosage of L-Orn, respectively. The size of the bubble indicates the sample size of the study.

Fig. 3.

Summary of study duration and daily dose. The x- and y-axes indicate the duration of administration and the dosage of L-Orn (mg/person/day), respectively. The size of the bubble indicates the sample size of the study. A: The dose indicates L-Orn hydrochloride (LOHC) or L-Orn-α-ketoglutarate (LOKG). B: The dose indicates net L-Orn

Regarding LOHC (Fig. 3A), the highest dose was 14,000 (10,989 as a net dose of L-Orn, Fig. 3B) mg/person/day (single dose) (Bucci et al. 1990) and the longest duration of administration was 90 days (Morita et al. 2013). The largest sample size in the included studies using LOHC was 28 cases (Fukuda et al. 2018).

The study by Sugino et al. involved healthy subjects being administered LOHC at 2000 mg (34.4 mg/kg BW) daily for 7 days and then 6000 mg (103 mg/kg BW) daily in two equally divided doses (51.5 mg/kg BW) for 1 day (Sugino et al. 2008). No adverse events were observed during this test period (Fig. 4A).

Fig. 4.

Summary of study duration and dose per kg of body weight or one-time dose per kg of body weight The x- and y-axes indicate the duration of administration and the dosage of L-Orn, respectively. The size of the bubble indicates the sample size of the study. A: The dose indicates the net daily dose of L-Orn per kg of body weight. B: The dose indicates the one-time dose per kg of body weight

Among the studies included in this SR, there were two trials whose aim was to evaluate the safety of L-Orn. Both studies used LOHC as the test sample. The study by Morita et al., in which healthy subjects were administered a single dose of LOHC at 3000 mg for 90 days, reported that no adverse events were observed during this test period (Morita et al. 2013). Miura et al. also evaluated the safety of LOHC in an escalation study in 23 healthy subjects, administering LOHC at doses of 6000, 9200, and 12,000 mg per day for 28 days each. They reported no L-Orn-related adverse events as a result (Miura et al. 2023). In the single-dose studies, the study by Bucci et al. reported gastrointestinal distress including cramping and diarrhea at a single dose of 170 mg/kg BW/time (10,989–14,000 mg/person) (Bucci et al. 1990). At lower doses (100 and 60 mg/kg BW/time), no gastrointestinal symptoms occurred (Fig. 4B).

There were three studies with LOKG, including two repeated-dose studies and one single-dose study. Of these, one repeated-dose study was classified into Category A and was excluded from the NOAEL or LOAEL evaluation. In the other repeated-dose study, no adverse events were observed at 2000 mg/person/day (net 952.4 mg/person/day) for 156 days. In the single-dose study, no adverse events were identified at a dose of 10,000 mg/person (net 6400 mg/person/time). Both doses were below the highest doses in studies using LOHC.

There were three studies in which the form of L-Orn was not known from the information in the article, but all were classified into Category A and were excluded from the NOAEL or LOAEL evaluation.

From these results, in the repeated-dose studies, the observed NOAEL of LOHC in healthy subjects was estimated to be 12,000 mg/person (139 mg/kg BW) per day and 9403 mg/person/day as net L-Orn dose. In the single-dose studies, NOAEL and LOAEL were 100 and 170 mg/kg BW/time, respectively, and as net L-Orn, 78.4 mg/kg BW/time and 133.2 mg/kg BW/time.

Meta-analysis of adverse events

We conducted a meta-analysis on the risk of AEs using research based on randomized controlled trials from Categories B and C. The studies by Bucci et al. (1990), Morita et al. (2013), Cynober et al. (1990), and Miura et al. (2023) were excluded from this meta-analysis because they did not include a control group and thus comparative analysis was not possible. The analysis of adverse events was conducted using the incidence of gastrointestinal symptoms.

All reported AEs of subjective symptoms were gastrointestinal, namely, nausea, vomiting, abdominal pain, and diarrhea. Five studies were eligible for a meta-analysis of gastrointestinal symptoms, and there were no reported AEs of gastrointestinal symptoms in any of these studies (Fig. 5).

Fig. 5.

Difference in risk of gastrointestinal symptoms associated with L-Orn The gastrointestinal symptoms included nausea, vomiting, abdominal pain, and diarrhea

Certainty of evidence

We confirmed variation in the certainty of evidence from the assessment of the RoB of the 22 studies targeted for the SR. We investigated whether random sequence generation and allocation concealment had been reported appropriately due to the focus on the incidence of AEs in this research. Within the RoB results, the category of “other bias” was particularly prominent, but it was generally due to the studies receiving funding from companies involved in L-Orn production/sales. The gastrointestinal symptoms that we focus on in this work are considered relatively temporary or tolerable. Indeed, in some reports, it is described that the intervention continued until the end of the study even if AEs occurred. This supports the assertion that the gastrointestinal AEs were minor events. In addition, we considered that there were few reports related to AEs among the studies targeted here because the researchers had emphasized assessing the effectiveness, rather than the safety. This is understandable because L-Orn is considered to be safe as a food component. For that reason, we thought that there was a serious risk of bias in the outcome based only on reports gathered at this time and we judged that it was appropriate to get one level of grade down according to GRADE.

From the results of the meta-analysis on the frequency of gastrointestinal symptoms, I² was small (0%, P = 0.96) and we also confirmed that the CIs overlapped between each study. Therefore, we judged that there is no problem with “inconsistency of results” according to GRADE.

In terms of the inclusion criteria for participants applied in this SR, were healthy subjects. Therefore, all studies were considered to adhere to the PICOS that we set and we judged that there were no problems with “indirectness of evidence” according to GRADE.

Generally, when conducting meta-analyses, a total number of events of 300 or more is preferable when the outcome is a binary variable. However, in this meta-analysis on the risk of gastrointestinal symptoms being associated with L-Orn administration, the number of events is below 300 and the 95% confidence interval of RD includes 0. This raised very serious concerns about “imprecision,” being equivalent to two levels of grade down according to the criteria of GRADE. However, GRADE was devised primarily designed to assess effectiveness guidelines. In contrast, our research is focused on safety, so we considered that there might be some limitations associated with using this classification system directly here.

Reporting bias could not be determined due to the small number of studies used in the meta-analysis. In addition, the number of studies that could provide information on adverse events was small so such compiled information was considered insufficiently accurate.

From the assessment of the five items (study limitations, inconsistency of results, indirectness of evidence, imprecision, and reporting bias), we eventually judged the certainty of the evidence as being intermediate by considering that there was insufficient support for asserting a lack of problems regarding “study limitations” and “imprecision.”

Discussion

In the present study, we evaluated the safety of L-Orn for healthy subjects through a systematic review. Evaluable studies were all of the available interventional studies on L-Orn, and all study designs were included. Our safety assessment focused on L-Orn added to the usual diet based on the clinical study designs of the research reports collected in the SR. The observed NOAELs investigated as indicators of the occurrence of adverse events were 12,000 mg/person/day in the form of LOHC and 9403 mg/person/day as the net dose. Of all the reported human studies that evaluated the safety of L-Orn collected by SR, that by Miura et al. (2023) was the most substantial in terms of sample size, duration of administration, and dosage. We therefore estimated the maximum daily dose of L-Orn on the basis of the results of that paper.

In an open-label, repeated-dose study, Miura et al. (2023) tested the effects of repeated L-Orn ingestion for 28 days by increasing the daily dose up to 12,000 mg/person/day and reported no adverse effects on biochemical parameters, blood pressure, mental and sleep parameters, or general conditions including body weight and nutrient intake. Although several adverse events were recorded, they were considered to be incidental and to have no relationship to L-Orn ingestion since they did not appear to increase in frequency with increasing doses.

The results of safety evaluations of urea cycle amino acids other than L-Orn in humans have been reported for arginine, which is located upstream, and citrulline, which is located downstream of L-Orn. McNeal et al. conducted a study with 30 g/day arginine for 90 days in adults, and they concluded that a long-term safe level of dietary arginine supplementation is at least 30 g/day in adult humans (McNeal 2021). Miura et al. investigated the safety of L-citrulline supplementation at a dose of 24 g/day for 28 days in healthy male adults, they reported no-observed-adverse-effect-level (NOAEL) of citrulline supplementation determined to be 24 g/day (Miura et al. 2023).

No adverse events were reported in the other studies reported in the articles collected in this SR, with the exception of one study by Bucci et al. (1990). They reported adverse events with strong causal relationships upon a single oral dose of L-Orn at 9900 mg/person, although these events were mild to moderate gastric cramps and diarrhea. This dose was lower than the tolerable dose (= 12,000 mg/person/day) in Miura’s study, as described above. As the adverse events reported by Bucci et al. were gastrointestinal symptoms, they may have occurred because the single dose was too high. Nonetheless, the study by Bucci et al. did not consider the possibility the question of dividing a single, large, and potentially effective oral dose into several smaller ones to be taken over the course of a day.

In contrast, based on confirmation that we were able to obtain from the authors, 12 g/man/day (171.9 mg/kg BW) of LOHC had been given in three separate doses per day in the study by Miura et al. Therefore, that study would have used a lower single dose (~ 4000 mg) than that of Bucci et al., which could explain the difference in tolerated daily dose between the two. That single dose is less than the largest single dose of 6.4 g/man/day (87.7 mg/kg BW) of LOHC among the studies included in this systematic review (i.e., in the work of Cynober 1990) that did not cause any adverse events. Thus, the maximum safe daily intake obtained from a single-dose study in this SR is 6.4 g/man/day (87.7 mg/kg BW) (Cynober 1990) in the form of LOHC. Meanwhile, the safe daily intake from repeated-dose studies is 12,000 mg/person/day (171.9 mg/kg BW) of LOHC when administered in three divided daily doses (Miura 2023).

L-Orn was found to have been supplied in three forms: L-Orn hydrochloride (LOHC), L-Orn aspartate (LOLA), and L-Orn α-ketoglutarate (LOKG). Since these three types of L-Orn are different chemical substances, if an adverse event occurs, it would be difficult to determine whether the cause is L-Orn or a counterpart substance if all studies were analyzed together. Therefore, all analyses were performed separately for each of the three chemical forms. Of the 18 studies (10 articles) corresponding to Categories B and C, where the presence or absence of adverse events can be determined, 16 were on LOHC. The remaining 2 studies were on LOKG, while none was on LOLA.

Since the purpose of this SR is to estimate safe daily intake, information obtained from repeated-dose studies and single-dose studies should be interpreted differently. Information on safe daily intake should be based on data obtained from repeated-dose studies. Meanwhile, single-dose studies are analyses of acute reactions and would provide information on safe doses on a “per dose” basis.

The present study has various limitations, which should be mentioned here. For example, there were small numbers of randomized controlled trials (5 studies) in Categories B and C. Therefore, further results regarding safety in randomized controlled trials are desired. In addition, the number of articles that could be evaluated regarding the occurrence of adverse events was small (10 articles). There were 13 articles that do not describe the occurrence of adverse events (i.e., Category A), so adverse events may have occurred. There is thus a need for further reports on assessments of the safety of L-Orn.

Conclusion

We assessed the safety of orally administered L-Orn by a systematic review based on trials on healthy humans. The main observed adverse events were gastrointestinal symptoms. Based on whether such adverse events occurred, for LOHC as the particular form of L-Orn, the observed NOAEL was 12,000 mg/person/day for healthy individuals. Besides, from the meta-analysis on the frequency of occurrence of AEs, the difference between those with and without L-Orn supplementation in the risk of gastrointestinal symptoms was 0.00 (95% CI: ±0.02, P = 1.00), indicating the absence of any significant differences caused by L-Orn.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Author contributions

HY performed the literature search, RoB assessment, and meta-analysis and drafted the article.YK performed the RoB assessment and GRADE assessment.SS performed the literature search, assessment of publication bias, and RoB assessment.RS provided overall supervision of the amino acid study and safety assessment.KH designed the study, provided overall supervision including RoB assessment, and drafted the article.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval

This study involved an SR and therefore was not reviewed by the Human and Animal Ethics Committee. The authors ensured that the work described in this manuscript complied with the ethical rules applicable to this journal.

Competing interests

The authors declare no competing interests.