Premature infants with BPD-PH are an understudied group of patients desperately needing new therapies. The current use of PH-targeted medications in neonates with BPD-PH is off-label. RCTs are needed to evaluate the safety and efficacy of PH-targeted pharmacotherapies in premature infants with or at risk of developing BPD-PH. Prior to performing RCTs, PK studies need to be conducted to guide the choice of doses to be evaluated for safety and efficacy [13]. As a first step in developing oral L-citrulline as a potential therapy for BPD-PH, we recently completed a study that simulated the PK profile of a multi-dosing regimen from data from a single dose of enterally administered L-citrulline in premature infants at risk of developing BPD-PH [17]. The current study showed that when six subjects were given even doses of L-citrulline four times a day, only two subjects (33%) achieved the target trough plasma L-citrulline concentrations of 50–80 μmol/L (8.76–14 μg/ml). Moreover, only one subject (17%) achieved a trough plasma L-citrulline concentration that increased 50–100% above baseline.

Several previously published findings influenced our decision to target steady-state trough L-citrulline plasma concentrations of 50–80 μmol/L in this study. We were aware that investigators interested in developing L-citrulline as a treatment to prevent post-operative PH in infants and children with congenital heart disease undergoing cardiopulmonary bypass had identified plasma L-citrulline concentrations above 37 μmol/L as being protective [20, 21]. A recent study also found that premature infants with BPD-PH had median plasma L-citrulline concentrations of 21 μmol/L, whereas median plasma L-citrulline concentrations were higher, 36 μmol/L, in infants with BPD who did not develop PH [22]. Findings from these studies suggest that achieving plasma L-citrulline concentrations above 37 μmol/L might be needed to prevent various forms of PH in children. Moreover, oral treatment with L-citrulline inhibited PH development in chronically hypoxic piglets who achieved 50–100% increases from their basal plasma L-citrulline concentrations [15, 16]. This latter finding led to our choice to target steady-state trough L-citrulline plasma concentrations at least 50–100% above basal levels of our patient population, using the median basal plasma L-citrulline concentrations of patients in our previous study, 31 μmol/L. Notably, the basal L-citrulline plasma concentrations of the 6 patients in this study, median 37.5 μmol/L, were similar to those in our previous study and comparable to previously published basal levels in pediatric patients [23].

The dosing strategy for this study was based on an optimal design dosage simulation-based methodology from our single-dose pharmacokinetic study [17, 24]. Simulations in our previous study predicted that enterally administering a L-citrulline daily dose of 150 mg/kg/d given once daily to eight times a day would achieve steady-state target trough plasma L-citrulline concentrations of 50–80 μmol/L (8.76–14 μg/ml) in premature infants at 32 ± 1 week postmenstrual age. In order to achieve trough plasma L-citrulline concentrations in the range of 80 μmol/L (14 μg/ml), doses of approximately 240 mg/kg/d are predicted to be needed. We considered practical issues when deciding the dosing frequency. The volume of each dose needed to be small enough to minimize the potential for emesis. In addition, since the long-range intent is that the L-citrulline will be administered for weeks to months, the frequency of dosing needs to not be overly burdensome for the caregiver. Therefore, a total daily dose of 240 mg/kg/d (total daily volume of 4.8 ml/kg/d) was divided into 60 mg/kg doses q 6 h (1.2 ml/kg/dose given 4 x a day) which considered both the desired target trough L-citrulline plasma concentrations and practical considerations of drug administration.

Of note, L-citrulline is endogenously produced from amino acids, including glutamine and proline, by enterocytes in the proximal intestines [25]. It is possible that differences in total daily amounts of dietary protein administered and absorbed by the gastro-intestinal tract might have contributed to the range in patient plasma L-citrulline concentrations measured both at baseline and after 72 h of L-citrulline administration. L-citrulline is considered a non-essential amino acid, so L-citrulline is not present in either infant formulas or parenteral nutrition. In addition, the normal human adult diet contains almost no L-citrulline, so the L-citrulline content in human milk is negligible [26]. Thus, other than the L-citrulline administered every 6 h as part of this study, no patient should have received an exogenous source of L-citrulline that would have influenced their L-citrulline plasma concentrations.

Orally administered L-citrullline has been shown to be highly bioavailable in adults [27]. L-citrulline is absorbed by the intestines via a number of amino acid transporters and then passes through the liver without major metabolism to reach the systemic circulation [25, 26]. The main organ of circulating L-citrulline consumption is the kidney [25, 26]. Approximately 75% of L-citrulline is converted by kidney proximal tubules into arginine, which is subsequently released into the renal vein and the systemic circulation [25, 26]. Some circulating L-citrulline is transported into vascular endothelial cells, including pulmonary arterial endothelial cells, by neutral amino acid transporters and then converted by a two-step enzymatic process into arginine [14, 26]. Less than 1% of an enteral dose of L-citrulline is excreted in the urine [27]. The potential impact from developmental changes on any of the preceding aspects of enterally administered L-citrulline processing is not yet known. However, it should be considered that variability in the bioavailability of the enterally administered L-citrulline, i.e., variability between patients in the amount of L-citrulline absorbed by the gastro-intestinal tract, may have contributed to the variability in degree of increase from baseline L-citrulline plasma concentrations and help explain why the trough increased 50–100% above baseline in only one of the six patients (17%).

Since a sizable amount of circulating L-citrulline is converted in the kidney to L-arginine, it is unsurprising that for most of the patients in the study the change in plasma concentrations of L-arginine paralleled those for L-citrulline. However, unlike L-citrulline, L-arginine is a semi-essential amino acid that must be provided as part of the human diet [28, 29]. This is particularly true for premature infants with a limited ability to produce L-arginine endogenously and reliance on exogenous, i.e. dietary sources, to maintain plasma L-arginine levels that are considered adequate for their needs [29]. Since the volume of human milk feedings ranged from 132–150 ml/kg/d, amounts of exogenously supplied dietary L-arginine might have contributed to the variability in plasma L-arginine concentrations between subjects in this study.

Oral administration of L-citrulline has been shown to increase the rate of plasma NO production in both children and adults [30, 31]. Urinary nitrites and nitrates have been used to reflect systemic NO production [32, 33]. Therefore, it may seem surprising that only two patients in this study had a change in urine NOx/Cr that paralleled increases in plasma concentrations of L-arginine and L-citrulline. Notably, other investigators have been unable to detect significant changes in urine nitrites and nitrates without administering very high doses of oral L-citrulline to adults (3 g L-citrulline twice a day for a week) [34]. Moreover, although rates of NO production increased with L-citrulline administration, plasma NOx concentrations did not change [30]. Both plasma and urine nitrite and nitrate concentrations are known to be confounded by dietary sources, limiting the ability of these measurements to reflect systemic or pulmonary NO production. Consequently, urine NOx/Cr measurements should not be relied on to accurately reflect changes in systemic or pulmonary NO production. Thus, even though we did not find an increase in Urine NOx/Cr with L-citrulline treatment in our group of 6 patients, it should not be concluded that systemic or pulmonary NO production did not increase with L-citrulline treatment in any or all the subjects.

Importantly, this study showed that this fragile patient population well tolerated the L-citrulline dosage strategy used in this study. For example, none of the patients developed evidence of gastrointestinal intolerance with the L-citrulline dosage strategy. Nor did any of the patients experience a decrease in systemic BP that warranted any therapeutic intervention. In addition, with exception of one subject who had a slight increase in respiratory support, the respiratory support was unchanged or reduced during study drug administration. This latter finding is important because neonates with respiratory diseases are at risk of developing pulmonary edema when total daily fluid volumes are increased. Maintaining respiratory stability in our patients in the face of the additional fluid volume required to administer the L-citrulline, 4.8 ml/kg/d in this case, is an important consideration.

We intended to study a population of premature infants at risk of developing BPD-PH. Premature infants born at ≤ 28 weeks gestation are known to be at greater risk of developing BPD than those born at more mature gestational ages [35, 36]. Current definitions of BPD are based on the respiratory support needs at 36 weeks postmenstrual age, with the definition of no BPD being the lack of respiratory support at 36 weeks of postmenstrual age [35, 36]. Therefore, infants were eligible if they had been born at < 28 weeks gestation and required respiratory support on the day of and for 14 days prior to study initiation. We chose to study infants at 32 ± 1 weeks postmenstrual age and not wait until 36 weeks, the postmenstrual age at which BPD is diagnosed. This choice is because our ultimate goal is to perform a RCT to determine whether starting oral L-citrulline treatment at or before 32 weeks postmenstrual age will reduce the percentage of neonates with BPD who have evidence of PH when they are ≥ 36 weeks postmenstrual age.

This study has limitations. One limitation is the small number of patients that we could enroll. Future Phase II studies need to be conducted in a larger number of patients to more completely evaluate the dosing strategies needed to achieve targeted steady-state plasma concentrations. Consistent with the experience of other investigators [37], we believe that one obstacle to successfully enrolling fragile patients into phase I studies like ours is that their parents were reluctant to grant consent for a study that had no direct benefit. Another issue is that none of our patients required invasive respiratory support at the time of study. Thus, our findings may not reflect patients who develop severe BPD. Another limitation is that because we performed the study with premature infants, we used a limited sampling strategy and collected only two blood samples per patient. The total volume of blood sampled during a study in newborns is limited in accordance with IRB guidelines. Meeting these guidelines often necessitates the use of sparse sampling strategies, which poses a major limitation to investigators evaluating pharmacokinetics of drugs in newborns. Limitations in the assays used to determine plasma L-citrulline and L-arginine concentrations may also have impacted our findings. We were also limited by not knowing the actual steady-state L-citrulline concentrations that will achieve therapeutic efficacy. Nor did we include efficacy end points. Indeed, it should be emphasized that no phase III RCT has been performed to evaluate and prove that targeting the NO pathway with any medication, including L-citrulline, is an efficacious therapy for infants with BPD-PH. This phase I study aimed to provide information to better inform the design of future phase II and phase III RCTs.

In summary, this study builds on the data from our previous single-dose PK study conducted in premature infants at risk of developing BPD-PH [17]. Although only two of the six subjects achieved our target trough L-citrulline plasma concentrations, these data can inform the choice of plasma sampling times and doses needed to achieve target L-citrulline plasma concentrations in a future phase II RCT performed in a larger number of subjects. Ultimately, phase III RCTs must be conducted to evaluate whether using an oral L-citrulline dosing strategy that achieves steady-state trough plasma L-citrulline concentrations of 50–80 μmol/L effectively prevents or ameliorates BPD-PH. Prior to performing a phase III RCT, a phase II RCT should be performed to evaluate the safety and provide some pharmacodynamic and preliminary efficacy information. Taken together, the results of this and our previous study provide the critical information needed to design a phase II RCT and take the next step towards evaluating the use of oral L-citrulline as a potential treatment to inhibit BPD-PH in premature infants.