HbSS affects one in every 400 African Americans(1). Children with HbSS often are of smaller weight and muscle mass when compared with unaffected black children and experience delayed growth and puberty(2, 3). These findings collectively are suggestive of a disease-specific disturbance(s) in whole body protein and/or energy metabolism. Accelerated body protein turnover and energy expenditure have been reported in adults with HbSS when compared with control subjects(4, 5). Such increases could be related to the increased energy required for increased red blood cell production as a result of the chronic hemolytic process and/or the increased cardiac output secondary to the chronic anemia. In growing children, disturbances of energy and protein metabolism are frequently manifested as delayed physical growth and development as is the case for HbSS in children.

Glutamine is a preferred fuel for tissues characterized by rapid rates of cell replication such as reticulocytes(6, 7) and serves as an important source of nitrogen for synthesis of purines and pyrimidines required in rapidly dividing cells(8). Animal studies suggest that maintenance of the free glutamine pool may be critical in the regulation of protein homeostasis because muscle glutamine pool size strongly correlates with rates of muscle protein synthesis(9, 10). In humans, increased efflux of glutamine from skeletal muscle leads to a precipitous depletion of muscle free glutamine pool during severe protein catabolic states(1113) and supplementation with parenteral glutamine decreased the incidence of sepsis and shortened hospital stay of patients receiving bone marrow transplantation for hematologic malignancies(14). In addition, glutamine is readily converted toα-ketoglutarate which is a key intermediate in the Krebs cycle(6, 15).

The present studies used indirect calorimetry and infusion of stable isotope-labeled amino aids to determine: 1) whether energy expenditure and protein turnover are increased in children with HbSS and2) whether an increased demand for glutamine as a result of increased protein turnover and energy requirements could lead to glutamine depletion, thus adversely affecting whole body protein metabolism in children with HbSS disease.


Materials. Solutions of L-[2-15N]glutamine (96.7%15 N), L-[1-13C]leucine (98.2% 13C), D,L-[2,3,4-2H3]glutamine, D,L-[4,5,5,5,6,6,6-2H7]leucine, and D,L-[4,5,5,5,6,6,6-2H7]KIS (Tracer Technologies (Somerville, MA) were prepared in 0.9% saline, tested for chemical, isotopic, and optical purity by gas chromatography-mass spectrometry, and verified to be sterile(plate culture) and pyrogen-free (Limulus lysate assay). Infusates were passed through a 0.22-μm Millipore filter and stored in sterile containers at 4 °C for <24 h until used.

Subjects. Written consent was obtained from the parents (and assent from the children) of 15 black children (8 patients with HbSS and 7 healthy children), according to procedures approved by the Nemours Children's Clinic Research Committee and the Institutional Review Committee at Baptist Medical Center, Jacksonville, FL.

The day before isotope infusion, weight, height, and pubertal status were recorded, and blood was obtained for Hb concentration and electrophoresis, complete blood count, reticulocyte count, sedimentation rate, and serum albumin. Patients with any evidence of infection, vaso-occlusive crisis, or intercurrent disease were excluded.

A dietary history was obtained before the study, and the patients were instructed to maintain their usual intake for the 5-7 d before isotope infusion. Body density was calculated based on the sum of skinfold thickness measured at four sites by the same observer, using a precision caliper(Holtain Ltd., Grymmych, Wales, UK), and the percent of body weight accounted for by fat was derived from body density(16). Lean body mass was estimated by the difference between body weight and fat mass(16).

Infusion protocol. The night before the study day, each subject ate dinner at 1800 h and a snack at 2000 h, and then remained fasting (with the exception of water ad libitum) until completion of the infusion study at 1200 h the following day.

On the next morning at 0700 h, each child was studied as an outpatient in the Clinical Investigation Unit at the Nemours Children's Clinic. Two short catheters were placed, one in a forearm vein for isotope infusion, and the other one in a superficial vein of the contralateral hand; the hand was placed in a warming pad at 60 °C to obtain arterialized-venous blood samples(17). Rates of Vco2 and Vo2, the respiratory quotient, and REE were determined by indirect calorimetry using a ventilated canopy(18). Breath aliquots were obtained for determination of 13CO2 enrichment just before the start, and during the last hour, of the isotope infusion(19).

Four-hour unprimed, continuous infusion of L-[2-15N]glutamine(0.18 μmol·kg-1·min-1), and primed, continuous infusion of L-[1-13C]leucine (5.6μmol·kg-1; 0.09μmol·kg-1·min-1) were administered in the postabsorptive state from 0800 to 1200 h.

One milliliter of arterialized-venous blood was sampled at 20-min intervals between 0 and 140 min to assess the rise of [15N]glutamine enrichment toward plateau(12, 14) and 2 mL at 160, 180, 200, 220, and 240 min for determination of steady state isotopic enrichment in KIC, leucine, and glutamine. Additional blood samples (10 mL) were obtained at 120, 180, and 240 min for determination of plasma creatinine, HCO-3, amino acids, and estradiol/testosterone concentrations. The total amount of blood sampled was therefore 50 mL, i.e. 2% of the total blood volume (80 mL/kg) of a 30-kg child.

Echocardiography. Two-dimensional echocardiograms(20) were performed on 12 of the patients (6 control subjects and 6 with HbSS) within 1 mo of the metabolic studies. Studies were performed with subjects in a resting state in a left lateral decubitus position, using either an Acuson 128 or Interspec Apogee CX imaging system. Screening two-dimensional study ruled out wall motion or valvular abnormatlities (other than trivial tricuspid or pulmonic regurgitation). Derived M-mode measurements were recorded from a parasternal short axis location at the level of the papillary muscles. Echocardiograms were reviewed by one of the authors (E.B.) without the knowledge of subject status.

Analytical methods. Known amounts of[2H3]glutamine, [2H7]leucine, and[2H7]KIC were spiked into each 100-μL aliquot of plasma to serve as an internal standard for measurement of glutamine, leucine, and KIC concentration by reverse isotope dilution. Plasma glutamine was isolated and derivatized as described elsewhere(21). KIC was extracted from 100 μL of plasma by passing the acidified plasma sample over an AG50 cation exchange column. To each KIC-containing fraction, 3 drops of 10 M NaOH and 200 μL of 0.36 M hydroxylamine HCl were then added, and samples were incubated at 60 °C for 30 min to produce an oxime derivative. Samples were then cooled immediately on ice, acidified with 2 M HCl, and spiked with 1 mL of supersaturated ammonium sulfate. KIC was then extracted twice by shaking after adding 8 mL of ethylacetate. The supernatant was then dried under nitrogen. Each dry sample was then spiked with 50 μL ofN- methyl-N-(t-butyldimethylsilyl)-trifluoroacetamide and incubated for 24-36 h at room temperature to obtain an oxime-t-butyldimethylsilyl KIC derivative. This modified method enhanced the sensitivity of the KIC assay, allowing for the use of smaller volumes of plasma than that of the previously describedt- butyldimethylsilyl derivative(22).

Isotopic enrichments in plasma leucine, KIC, and glutamine were determined by selected ion monitoring gas chromatography-mass spectrometry(Hewlett-Packard MSD 5970) as described previously(21, 22). For KIC, ions at m/z = 316, 317, and 323, representing the prominent ions of natural KIC, [13C]KIC, and [2H7]KIC, respectively, were selectively monitored.

Breath 13CO2 enrichments were measured by isotope ratio mass spectrometry (Isochrom III, VG, Ipswich, UK). Plasma amino acid concentrations were determined using a Beckman 7300 amino acid analyzer (Beckman Instruments, Palo Alto, CA), and glucose concentration, by using the glucose-oxidase method.

Calculations. Glutamine appearance rate(RaGLN, μmol·kg-1·min-1)-which is synonymous to the glutamine utilization rate under conditions of steady state-was calculated as: RaGLN =i[(Ei/Ep) - 1], where i is the tracer infusion rate(μmol·kg-1·min-1) andEi and Ep are the stable isotope enrichments (mol% excess) in the infusate and in plasma glutamine at isotopic steady state, respectively(19, 23, 24).

Leucine Ra (RaLEU) was calculated using an analogous equation where Ep is plasma[13C]KIC enrichment(19, 23, 24). Leucine oxidation (Ox, μmol·kg-1·min-1) was calculated from the following equations: whereE co2, Ep, andEi are the steady state 13C enrichments(mol% excess) in expired CO2 (as determined from samples obtained over the last hour of isotope infusion), plasma KIC, and the infused leucine, respectively; wt (kg) is the patient's body weight; 44.6 converts mL CO2/min to μmol/min; 0.81 is the estimated fraction of CO2 recovered in expired air; and 100 converts enrichments (%) to fractions of unity. NOLD, an estimate of leucine incorporation into protein, was calculated as NOLD = RaLEU - Ox(19, 23, 24).

Both glutamine release from protein breakdown (BGLN) and glutamine de novo synthesis (DGLN) contribute to the appearance of glutamine, a nonessential amino acid. BGLN was estimated as BGLN = b ×RaLEU, where b is the ratio of glutamine to leucine content of body protein, based on the following assumptions: 1) the release of an amino acid from protein breakdown is proportional to its abundance in whole body protein; 2) in the postabsorptive state, protein breakdown is the only net source of leucine, an essential amino acid, and body protein contains 8 g of leucine/100 g of protein(19); and 3) although its exact abundance remains to be defined, glutamine contributes half the 13.9 g/100 g of protein known to represent the total glutamine + glutamate content of protein(25). It follows that: DGLN =RaGLN - BGLN(25).

The rise of plasma [15N]glutamine enrichment from 0 to 240 min was fitted to a single exponential curve using nonlinear regression(23, 24): EGLN =EpGLN [1 - e-kt], whereEGLN is a function of sampling time (min). The fitted parameters are enrichment at plateau, EpGLN (mol% excess), and the rate constant for glutamine turnover, k (min-1). Tracer-miscible glutamine pool size (poolGLN) can be calculated asRaGLN/k for each individual subject.

Left ventricular volumes were calculated from M-mode data using the Teicholz method(20). Left ventricular mass was calculated using the American Society of Echocardiography “leading edge” method(26) and was corrected for both body surface area and body mass index. Left ventricular volume/mass ratio was calculated according to the method of Byrd et al.(32), as modified by Silverman(20). Normal for this ratio is 0.97 ± 0.16(20).

Statistical analysis. Data are presented as means ± SE. During the tracer infusions, steady state for Vco2 and plasma amino acid levels and enrichments were defined by the absence of a significant correlation of the measured parameter versus time over the considered period. Data were compared using a 2-tailed unpaired t test.


Table 1 summarizes relevant clinical characteristics of the populations. Patients and controls were all Tanner I or II and well matched for age. All girls were Tanner I, based on physical examination and serum estradiol concentrations (0.8 ± 0.3 versus 0.6 ± 0.2 pg/mL in HbSS versus controls, respectively; NS); among boys, mean serum testosterone was 16 ± 13 versus 22 ± 10 ng/dL in HbSS versus controls; one HbSS subject and two control subjects had serum testosterone concentrations consistent with early Tanner II stage. Sex ratio (male/female) was 4/4 in HbSS versus 5/2 in control subjects. Indeed, most healthy 10-13-y-old girls have already entered puberty, and we experienced more difficulty recruiting prepubertal girls in the desired age range. Although there was a trend toward a lower height and weight in the sickle cell patients group, the difference failed to reach statistical significance. Body composition was also similar between the two groups.

Table 1 Relevant clinical characteristics of the population

When expressed on a whole body basis, REE did not differ between groups: 1367 ± 48 versus 1283 ± 63 kcal/d. However, when REE was expressed per unit of body weight, it was 27% higher in HbSS children than in controls (49 ± 3 versus 38 ± 1 kcal·kg-1·d-1); a 19% difference persisted when REE was expressed per unit of lean body weight (56 ± 3versus 48 ± 2 kcal·kg lean body mass-1·d-1 (both p < 0.01;Fig. 1).

Figure 1
figure 1

Rates of energy expenditure (expressed in kcal·kg-1·d-1, left; and in kcal·kg lean body mass-1·d-1, right) in eight children with sickle cell anemia (solid bars) and seven age-matched controls (open bars). Values are mean ± SE;* statistically significant difference between groups.

Plasma leucine (76 ± 4 versus 70 ± 6 μM), KIC(25 ± 2 versus 21 ± 1 μM), and glutamine (522± 47 versus 508 ± 31 μM) concentrations did not differ between the groups. Children with HbSS had higher plasma asparagine (37± 1 versus 31 ± 2 μM; p < 0.05) and cysteine (36 ± 3 versus 19 ± 3 μM; p < 0.02) concentrations than controls; none of the other 18 amino acids measured differed significantly between groups. As shown in Fig. 2, leucine Ra was 58% higher in sickle cell subjects than in controls: 3.67 ± 0.21 versus 2.32 ± 0.18μmol·kg-1·min-1, respectively (p < 0.001). Leucine oxidation was slightly, but not significantly higher in HbSS subjects than in controls: 0.38 ± 0.09 versus 0.32 ± 0.03 μmol·kg-1·min-1 (NS). When one outlyer patient (whose oxidation was >3 SE above the group's mean) was excluded, leucine oxidation was, in fact, identical in both groups: 0.32 ± 0.07versus 0.32 ± 0.03μmol·kg-1·min-1 in HbSS versus control children, respectively. NOLD, was 65% greater in HbSS subjects than in controls: 3.29 ± 0.18 versus 1.99 ± 0.16μmol·kg-1·min-1 (p < 0.001). Both leucine Ra and NOLD were correlated with Hb (r = 0.907 and 0.936, respectively; p < 0.001), and reticulocyte count(r = 0.829 and 0.816, respectively, p < 0.001;Fig. 3).

Figure 2
figure 2

Leucine (top) and glutamine(bottom) kinetics in eight children with sickle cell anemia(solid bars) and seven age-matched controls (open bars).Ox, oxidation; DGLN, de novo glutamine synthetic rate(μmol·kg-1·min-1); Pool, tracer-miscible glutamine pool size (μmol·kg-1). Values are mean ± SE; *statistically significant difference between groups.

Figure 3
figure 3

(Top) Correlation between leucine appearance rate (Ra, μmol·kg-1·min-1), and rates of energy expenditure (kcal·kg-1·d-1).(Bottom) Correlation between reticulocyte count (103/μL) and leucine Ra, in eight children with sickle cell anemia(solid circles) and seven age-matched controls (open circles)

Glutamine Ra was 47% increased in sickle cell children(8.41 ± 0.79 versus 5.73 ± 0.29μmol·kg-1·min-1; Fig. 2). This was accounted for by increases in both 1) glutamine de novo synthesis (6.9 ± 0.8 versus 4.9 ± 0.3μmol·kg-1·min-1 in HbSS versus control children respectively; p < 0.05), and 2) release of glutamine from protein breakdown (2.1 ± 0.1 versus 1.3 ± 0.1 μmol·kg-1·min-1; p< 0.001). The size of glutamine's tracer-miscible pool was determined in seven control subjects and six patients; although there was a trend toward a greater pool size in HbSS children than in controls, the difference failed to reach statistical significance, whether expressed per unit of body weight (461± 85 versus 303 ± 21 μmol·kg-1; NS) or of lean body mass (502 ± 108 versus 377 ± 32μmol<kg LBM-1; NS). Indeed, when two outlyers who were at >2 SE above the group's average value were excluded from the analysis, pool size was identical in both groups: 338 ± 54 versus 303 ± 21μmol/kg (NS) in HbSS versus control children, respectively. Glutamine Ra was correlated with reticulocyte counts (r = 0.756; p < 0.002).

REE was significantly correlated with rates of leucine Ra(r = 0.716; p < 0.005; Fig. 3), nonoxidative leucine disposal (r = 0.704; p < 0.005), glutamine Ra (r = 0.742; p < 0.002), and glutamine de novo synthesis (r = 0.667; p < 0.01). REE was significantly correlated with reticulocyte count (r = 0.736; p < 0.002) and negatively correlated with Hb concentration(r = -0.585; p < 0.05).

Echocardiography (Table 2) revealed significant differences in resting cardiac index and left ventricular mass (p< 0.01), left ventricular volume/mass ratio (p < 0.05), as well as in left ventricular end-diastolic dimension, interventricular septal thickness and left ventricular posterior wall thickness, as measured in direct M-mode (p < 0.05), between groups. No significant differences in left ventricular mass was detected between male and female subjects. REE correlated with cardiac index (r = 0.668; p < 0.01).

Table 2 Echocardiographic measurements obtained in six children with HbSS and six control children


The present study demonstrates that rates of energy expenditure, protein turnover, and glutamine utilization, are increased in children with sickle cell anemia, when compared with age-, sex-, and race-matched control children, even in the absence of vasooclusive or sickling crisis, or intercurrent illness. Whatever the mechanism(s), the increased calorie and protein demands seem to be disease specific, and may warrant the recommendation of increased dietary allowances for children with sickle cell disease.

Although increased energy expenditure has previously been reported in 6 adult and adolescent HbSS patients(4, 5), the present study is first to document such an increase in prepubertal children with the disease. This finding cannot be explained by differences on body composition, undernutrition, or stress. HbSS patients indeed were slightly shorter and lighter (Table 1) than control subjects; smaller humans will invariably have a relatively larger metabolic rate without the presence of disease, when the measured parameters are expressed per unit of body weight, whereas they have values identical to larger individuals when data are normalized to metabolic body weight (weight0.75); for instance, healthy neonates have greater rates of leucine appearance than healthy adults(27). In the present study, as patients and control subjects were well matched for age, normalization of the measured parameters to lean body mass, a commonly used index of metabolically active mass, failed to abolish the difference between groups. Similarly, when REE was expressed in kcal·kg-0.75·d-1, it was still21% higher in sickle cell patients, compared with controls (112 ± 6 versus 92 ± 3 kcal·kg-0.75·d-1 in sickle cell children and control subjects, respectively; p < 0.05). Moreover, decreased-rather than increased-energy expenditure is characteristic of malnutrition(28). Besides, because none of the HbSS children studied had any evidence of intercurrent illness, nor sickling or vaso-occlusive crisis, acute stress may not account for our findings. Increased production of erythrocytes could contribute to the higher REE. Indeed, reticulocyte count accounted for 54% of the variance(r2 = 0.542) in REE when both the HbSS and the control groups were analyzed together.

Chronic anemia results in increased cardiac work load. Resting cardiac index and left ventricular mass were increased in our patients(Table 2). This is consistent with previous studies in adults and children with sickle cell anemia(2933). Yet, in contrast to previous studies in adults with HbSS documenting an increased left ventricular volume/mass ratio(30), our children with HbSS showed a tendency toward a lower volume/mass ratio, more typically seen with pressure overload. This unexpected finding may reflect an “inappropriate” ventricular hypertrophy(34) and contribute to increased energy expenditure.

The kinetics of the essential amino acid leucine have been extensively used as a paradigm for whole body protein metabolism (e.g. seeRefs. 19, 24, 25, and 35). Proteolysis is the only source of leucine in the postabsorptive state, and the higher leucine Ra observed in children with HbSS in the present study suggests increased rates of protein breakdown.

Similarly, the increased rate of NOLD is suggestive of increased whole body protein synthesis. When rates of leucine oxidation are measured, it must be borne in mind that the appearance of 13CO2 at the mouth depends not only on the release of 13C from the labeled leucine during cellular metabolism, but also on the dynamics of the transport of 13CO2 and H13CO3 from cell to alveolar gas. In both groups, an apparent steady state had been reached over the last hour of isotope infusion in expired air 13CO2 enrichment, as attested by a mean coefficient of variation of 6.9 ± 2.4% and 7.9 ± 2.6% in breath13 CO2 enrichment over the 4th h of isotope infusion in sickle cell patients and controls, respectively, so that steady state equations were used to calculate leucine oxidation. Yet, in patients with chronic anemia, the combination of an increased cardiac output with a reduced blood bicarbonate pool may increase the respiratory excretion of metabolically produced13 CO2. The recovery of 13CO2 was not measured in our patients and was assumed to be 81% in both groups (see“Methods”), i.e. the same as documented in healthy adults(19). If we assume that the recovery of metabolically produced 13CO2 was increased to 100% in HbSS patients, compared with 81% in controls, this would decrease the calculated rate of leucine oxidation to 0.31 ± 0.07μmol·kg-1·min-1 in the patients, a value virtually identical to that of the controls (0.32 ± 0.03μmol·kg-1·min-1). Calculated NOLD would be slightly increased-to 3.36 ± 0.19μmol·kg-1·min-1 in HbSS children, yet the difference between groups would not be altered, and our conclusions would remain qualitatively similar.

Based on estimated red cell mass and erythrocyte life span, it was proposed that the average rate of Hb synthesis may be 0.725 g·kg-1·d-1 in adults with HbSS, compared with only 0.094 g·kg-1·d-1 in healthy adults(4). The 1.3μmol·kg-1·min-1 difference in NOLD observed in the current study between HbSS and controls, translates to a 3.05-g protein·kg-1·d-1 difference, assuming body protein contains 8 g of leucine/100 g of protein. Thus, increased Hb synthesis could account for only 20% of the difference. Yet these calculations do not include the synthesis of other structural erythrocyte proteins. Reticulocyte count indeed accounted for 69 and 67% of the variance in leucineRa and NOLD, respectively. Because both protein breakdown and protein synthesis are energy-requiring processes, the rise in leucineRa and NOLD undoubtedly contributes to the rise in energy requirement. Indeed, leucine Ra and NOLD accounted for 51 and 50% of the variance in REE, respectively. Because leucine oxidation was not significantly increased, protein wasting was not observed in our patients. This suggests that they were still in a compensated situation. Indeed, their protein intake seemed to be slightly-although not significantly-higher than the controls' intake (Table 1). Yet whether this intake meets the needs for normal growth remains to be determined.

We observed increased rates of whole body glutamine utilization in patients with HbSS (Fig. 2). We speculate that, in HbSS children, the shortening of erythrocyte life-span(1) and the consequent accelerated rate of cell division in the bone marrow may chronically increase the rate of glutamine utilization. This increased rate of glutamine utilization may, in the presence of inadequate intake due to anorexia and/or insufficient rates of glutamine de novo synthesis, lead to depletion of the body glutamine stores, which may contribute to a decreased rate of protein synthesis in skeletal muscle, and, as a consequence, to poor growth, if glutamine is indeed instrumental in the maintenance of protein homeostasis. We did not, however, observe any decrease in rates of glutamine de novo synthesis, nor any decrease in the size of the tracer-miscible glutamine pool. Thus-in this group of relatively well nourished HbSS children studied in the absence of any infection of vaso-occlusive event-the alterations in whole body protein turnover observed do not seem to result from glutamine depletion. We have shown that a large fraction (30-50%) of glutamine flux is oxidized in vivo in healthy adults(36). Glutamine utilization and oxidation are known to increase with stress. It remains to be determined whether glutamine utilization exceeds the endogenous capacity for glutamine de novo synthesis and results in glutamine depletion when children with sickle cell disease suffer from acute intercurrent illness.

Taken together, the results of the present study show that, even in the absence of sickling or vaso-occlusive crisis or intercurrent illness, prepubertal children with HbSS use 19% more energy, 58% more protein, and 47% more glutamine than their healthy counterparts. Increased bone marrow activity and increased cardiac output may contribute to this state of hypermetabolism. Regardless of the mechanism of protein and energy wasting, these findings emphasize the need to assess the calorie and protein intake of children with sickle cell anemia.

Consistent with the present findings, acceleration of growth was documented by other workers in two of three children with sickle cell anemia who received oral or nasogastric dietary supplementation(37). Prospective studies with dietary supplementation in larger groups of patients would be needed to demonstrate the potential benefit of nutritional supplementation in children with HbSS.