Steven P. Conway
ABBREVIATIONS. BMD, bone mineral density; BMC, bone mineral content; BSAP, bone-specific alkaline phosphatase; CF, cystic fibrosis; Gla-OC, carboxylated osteocalcin; Glu-OC, undercarboxylated osteocalcin; PICP, procollagen I carboxy-terminal propeptide; PIVKA-II, prothrombin produced in vitamin K absence; MAb, monoclonal antibody; OC, osteocalcin; AU, arbitrary unit.
Dietary vitamin K is available mostly in green leafy vegetables and some vegetable oils (eg, soybean, rapeseed, and olive oils), as phylloquinone (vitamin [.1]). (1) Gut bacteria synthesize menaquinones (vitamin [.2]), but it is unclear how much these contribute to overall vitamin K status. (1,2) Members of both series can function as the necessary cofactor for the posttranslational conversion of specific glutamyl residues to [gamma]-carboxyglutamyl residues in vitamin K-dependent proteins, such as prothrombin and osteocalcin (OC). The latter, which is produced by osteoblasts, accounts for 10% to 20% of noncollagenous bone proteins, has a high affinity for hydroxyapatite, and plays a regulatory role in bone formation and mineralization. Circulating OC levels are correlated with bone formation. (3)
Undercarboxylated vitamin K-dependent proteins are functionally defective, and their measurement in serum offers a sensitive marker of vitamin K function for the tissue in which the protein is synthesized. (4) The level of undercarboxylated prothrombin, ie, prothrombin produced in vitamin K absence (PIVKA-II), is a much more sensitive indicator of suboptimal hepatic stores than is the prothrombin time, which remains unchanged until the prothrombin concentration falls below 50% of normal values. (5)
The liver is more efficient than bone in using available vitamin K and, in the absence of malabsorption, it is rare for a nutritional vitamin K deficiency to be of sufficient severity to impair blood coagulation. The bone vitamin K-dependent proteins are more vulnerable, and a degree of undercarboxylation of OC is naturally present among healthy people of all age groups. (6-9) The measurement of the fraction of total OC as undercarboxylated OC (Glu-OC) is a sensitive indicator of the vitamin K status of bone. (4,8) It is the first functional marker to respond to vitamin K depletion and the last to respond to repletion. (10)
Children with cystic fibrosis (CF) are at risk of vitamin K deficiency as a result of fat malabsorption, bowel resection, liver disease, and chronic antibiotic use. Isolated cases of severe bleeding with hypoprothrombinemia in infancy were reported in the early 1970s (11,12) and may still occur when neonates with CF receive only oral vitamin K supplements. (13) Routine vitamin K administration at birth removes this risk. (14) Early studies of vitamin K status among older children and young adults with CF yielded conflicting results, documenting adequate vitamin K levels for most patients (15,16) but also vitamin K-associated coagulopathy (17) and increased PIVKA-II levels. (18) Fifty-six of 128 patients (44%) in those 4 studies (15-18) is were receiving vitamin K supplements. In a review article in 1994, Durie (19) concluded that there had not been adequate evaluation of patients at risk but only "conflicting and incomplete answers." Rashid et al (20) subsequently used a sensitive immunoassay for PIVKA-II to assess vitamin K deficiency among 98 pediatric and adult patients with CF who were not receiving vitamin K supplements. Seventy-eight percent of pancreatic function-insufficient patients and all patients with CF-associated liver disease had increased PIVKA-II levels. Similarly, Wilson et al (21) reported abnormal PIVKA-II levels for 81% of 72 pancreatic function-insufficient patients with a mean age of 15 years.
A limitation of using PIVKA-II as the only index of vitamin K status (20,21) is that it may not distinguish between impaired carboxylation of prothrombin attributable to nutritional vitamin K deficiency and that attributable to liver dysfunction. (22)
The aim of this study was to assess vitamin K status in an unselected population of children with CF and to investigate any relationship between vitamin K and bone mineral status and bone turnover. To the best of our knowledge, this is the largest study to document the prevalence of vitamin K deficiency in CF with a combination of tests, including serum phylloquinone measurements to assess tissue stores and specific and sensitive monoclonal antibody (MAb)-based immunoassays to assess the carboxylation of prothrombin in the liver and of OC in bone.
METHODS
Patients
Children [greater than or equal to] 5 years of age who were attending the Leeds pediatric CF unit were invited to take part in the study. One hundred six children (52 male) entered the study. The z scores for height, weight, and BMI were derived from the 1990 British reference standards. Fifteen children were receiving oral vitamin K supplementation at a dose of 10 mg/day. In accordance with the policy of the CF unit, all pancreatic function-insufficient patients received supplements of the fat-soluble vitamins A and D, with dosages adjusted according to serum levels. Fasting blood samples were obtained, usually between 8:30 and 9:00 AM.
Vitamin K studies, serum and urinary bone turnover marker levels, and serum levels of vitamin A (retinol) and vitamin D (25-hydroxyvitamin D) were obtained for most of the children (Table 1). One hundred six patients underwent dual-energy x-ray absorptiometric studies. Three patients had sufficient pancreatic function, and 52 had CF-associated liver disease, defined by the presence of hepatomegaly, increased serum hepatic enzyme levels, and/or abnormalities of the liver on ultrasound scans.
Assessments of Bone Mineral Density and Bone Turnover
Bone mineral density (BMD) and bone mineral content (BMC) were measured at the lumbar spine (L2-L4) (spinal BMD and BMC) and for the total body (total-body BMD and BMC) with a GE Lunar Prodigy densitometer (GE Lunar, Madison, WI). Age-matched z scores for BMD were derived from the reference ranges provided by the manufacturer; analysis was performed with software version . Bone formation and resorption were assessed by measuring plasma bone-specific alkaline phosphatase (BSAP) levels, procollagen I carboxy-terminal propeptide (PICP) levels, and free urinary pyridinoline, deoxypyridinoline, and N-telopeptide levels (second morning void urine sample). Urinary bone turnover marker levels were standardized against urinary creatinine valuses.
Assessments of Vitamin K Status
Serum phylloquinone (vitamin [.1]) concentrations were measured with high performance liquid chromatography with post-column reduction and fluorimetric detection, as described previously. (23) This method has a lower limit of quantification of ng/mL. The reference range for healthy normolipemic adults is to ng/mL (median: ng/mL) in the fasting state and to ng/mL (median: ng/mL) in the nonfasting state. We defined a suboptimal vitamin [.1] concentration as [less than or equal to] ng/ m L.
Serum PIVKA-II levels were measured with a conformation-specific MAb (designated C4B6) in an enzyme-linked immunosorbent assay format. (24,25) Results are expressed as arbitrary units (AUs) per milliliter because circulating PIVKA-II may represent multiple forms of partially carboxylated prothrombin and neither their relative abundance in plasma nor their relative affinity for the antibody is known. Calibration is therefore made against PIVKA-II species purified from patients treated with oral anticoagulant therapy. With electrophoretic techniques, 1 AU is equivalent to 1 [micro]g of purified PIVKA-II. The minimal limit of detection is ~ AU/mL, and we defined values of [greater than or equal to] AU/mL as indicative of an abnormally increased concentration of PIVKA-II.
Glu-OC and carboxylated OC (Gla-OC) levels were determined with separate immunoassays, in an enzylne-linked immunosorbent assay format, with Glu-OC and Gla-OC kits from Takara Shuzo (Tokyo, Japan). These assays use different MAbs that recognize fully decarboxylated or carboxylated OC, with little or no cross-reactivity. In line with previous recommendations for assessments of the vitamin K status of bone. (8) the Glu-OC fraction was also expressed as a percentage of total OC. A measure of total OC was evaluated by summing the Glu-OC and Gla-OC fractions.
Statistical Analyses
Statistical analyses were preformed with Minitab version (Minitab, State College, PA). Anderson-Darling statistical tests were used to determine whether the data were normally distributed. Because most of the data had a nonparametric distribution, results are given as medians and 95% confidence intervals. All patients with undetectable vitamin K levels were excluded from correlation analyses involving vitamin K. Similarly, patients without increased PIVKA-II levels were excluded from correlation analyses involving PIVKA-II. All other correlation analyses used the full data sets. Correlation analyses were performed with logarithmically transformed values for vitamin K and PIVKA-II levels, because of nonparametric distributions. Comparisons of z score results with zero reference values were performed with the 1-sample test. Spearman's rank correlation coefficients were used to evaluate the possible relationships between various parameters. A P value of <.05 was considered significant. The study was approved by the hospital ethics committee, and informed written parental consent was obtained for all study participants.
RESULTS
Patient characteristics and measured parameters are shown in Table 1. Mild growth impairment was observed for the patients with CF. Pulmonary function tests and Shwachman-Kulczycki scores indicated mild/moderate disease severity. Age-matched BMD z scores showed a small but significant reduction at the lumbar spine, which might be a reflection of body size. Although we had no control group of the same age range, the median serum vitamin [.1] concentration of ng/mL for the children with CF was greatly reduced, compared with the median of ~ ng/mL for healthy fasting adults. Fifty-four of the 93 children with vitamin K study results had either undetectable ([less than or equal to] ng/mL) or suboptimal ([less than or equal to] ng/mL) serum vitamin [.1] levels, suggesting low tissue stores (Fig 1). Thirty-nine children, 28 of whom also had low vitamin [.1] levels, had increased PIVKA-II levels (> AU/mL), suggesting a functional abnormality of [gamma]-carboxylation of prothrombin (Fig 2). Sixty-five of the 93 children (70%) therefore showed suboptimal vitamin K status, on the basis of low serum vitamin [.1] levels, increased PIVKA-II levels, or both abnormalities.
[FIGURES 1-2 OMITTED]
Of the 15 children prescribed vitamin K supplements, 4 had increased vitamin K levels (3 with normal and 1 with increased PIVKA-II levels), 2 had normal vitamin K and PIVKA-II levels, 2 had low vitamin K levels (1 with normal and 1 with increased PIVKA-II levels), and 7 had undetectable vitamin K levels (5 with increased and 2 with normal PIVKA-II levels). Poor adherence to therapy was suspected for 5 children, including 4 in the latter group. The median serum vitamin [.1] level for the children who received vitamin K supplements and had detectable levels was ng/mL (range: .68 ng/mL). Analysis of the 78 non-supplement-treated children showed that 48 had suboptimal or undetectable levels of vitamin K and 35 had increased serum PIVKA-II levels; 23 of the latter children also had low serum vitamin K levels.
Serum vitamin K levels were logarithmically transformed, and weak negative correlations were observed with Glu-OC levels (r = , P = .01) and percentages of Glu-OC (r = , P < .05). No other significant correlations were observed between vitamin K levels and bone turnover markers or bone parameters. Serum PIVKA-II levels were also logarithmically transformed, but no correlation was observed with any of the measured parameters.
Median levels of Glu-OC and Gla-OC were 43 ng/mL and ng/mL, respectively, whereas the percentage of Glu-OC was 71%. Both total OC (represented by the sum of Glu-OC and Gla-OC fractions) and Glu-OC levels were correlated positively with bone marker levels. The strongest correlation was observed with the bone formation markers BSAP and PICP. The percentage of Glu-OC was correlated with bone formation markers less strongly than Glu-OC levels and was not correlated with bone resorption markers. The only association observed with Gla-OC levels was a weakly significant correlation with the bone resorption marker free deoxypyridinoline/creatinine. No significant correlation was observed between Glu-OC levels, Gla-OC levels, or percentage of Glu-OC and any bone parameters (Table 2).
The measured bone parameters spinal BMD, spinal BMC, total-body BMD, and total-body BMC were correlated negatively with both bone formation and resorption markers. Slightly stronger correlations were observed with bone resorption markers. No correlations were observed with vitamin A or vitamin D levels.
DISCUSSION
The data show that vitamin K deficiency is common among children with CF. On the basis of well-validated assays of vitamin K status, we found that approximately one third of the children (28 of 92 children) had both low serum vitamin [.1] levels and increased PIVKA-II levels, whereas slightly more than one half (54 of 93 children) had suboptimal serum levels of the vitamin. Little is known about the interrelationship of these 2 markers as vitamin K deficiency progresses, but it is known that a rapid decline in hepatic vitamin [.1] stores among surgical patients is mirrored by rapidly declining serum vitamin [.1] levels. (26) After prolonged depletion, the hepatic stores of vitamin [.1] become insufficient to sustain [gamma]-carboxylation of the vitamin K-dependent coagulation factors, and PIVKA-II is produced and released into the circulation. Among healthy subjects, there is a weak but significant negative association between plasma vitamin [.1] levels and PIVKA-II levels. (27)
It is more difficult to explain the finding of increased PIVKA-II levels in the presence of normal serum vitamin [.1] levels, which was noted for 13 patients. Apart from the possible influence of recent meals, increased PIVKA-II levels may indicate a degree of hepatocellular injury, which may be mild. (25,28) A lack of exact correlation between serum vitamin [.1] levels and PIVKA-II levels (determined with the same assays) was found recently for ICU patients. (22)
There was no direct association between vitamin K levels and bone turnover or bone density. This may reflect several confounding factors, such as the impact of the dietary intake in the previous 24 hours on serum vitamin K levels. Another limitation was the fact that many children (33%) had undetectable circulating vitamin [.1] levels, which reduced the statistical power of finding an association. An intervention study is required to show definitively whether there is a correlation between serum vitamin K levels and bone turnover. Vitamin K levels showed a significant negative correlation with levels of Glu-OC, the precursor and presumed inactive form of OC. The latter was correlated significantly with levels of bone turnover markers, as were levels of total OC (71% of which was in the undercarboxylated form). Increased skeletal turnover is associated with rapid bone loss. (29) Through its role in the carboxylation of OC, vitamin K deficiency in CF may be associated with an uncoupling of the normal balance between bone formation and resorption. Vitamin K supplementation for patients with CF may be necessary to avert CF-associated premature osteopenia and osteoporosis. Vitamin K has an excellent safety profile, with no adverse effects being reported with administration of natural sources of the vitamin.
The molecular role of OC remains unclear, but studies in non-CF populations have provided evidence that improving the carboxylated fraction of OC has a positive effect on bone health and that vitamin K insufficiency contributes to the development of osteoporosis. (30) Low circulating vitamin levels have been documented among elderly women and men with reduced BMD or acute fractures, compared with control subjects. (31-33) Postmenopausal women showed an inverse relationship between Glu-OC levels and BMD. (34) Circulating levels of Glu-OC were increased significantly and were correlated with the subsequent risk of hip fracture. (35) In this patient group, the Glu-OC level was identified as an independent variable from which BMD could be calculated with an accuracy of >70%. (7) Among patients with longstanding Crohn's disease, vitamin K levels were significantly reduced and Glu-OC levels were high, compared with control subjects. (36) The Glu-OC level was also an independent risk factor for, and showed an inverse correlation with, low BMD at the lumbar spine.
Vitamin K supplementation may reduce bone loss. Among postmenopausal women, it increased the fraction of Gla-OC and reduced urinary calcium excretion and urinary markers for bone resorption. (37,38) Other intervention studies showed that vitamin K supplements for a variety of subjects could reduce Glu-OC levels, increase bone formation, (4) and improve bone health (32,39) significantly. Data from animal model systems support the clinical trials showing that vitamin K stimulates bone formation and reduces bone resorption. (4) Moreover, vitamin K may have a synergistic interaction with vitamin D. (40,41)
There is no established dietary reference value for vitamin K in the United Kingdom. Department of Health guidelines (42) suggest that a daily intake of 1 [micro]g/kg body weight is adequate for the coagulation function of vitamin K. Studies suggested that mean dietary intakes of phylloquinone among young and old adults in the United Kingdom were in the range of 65 to 85 [micro]g/day. (43,44) However, a substantial proportion of the United Kingdom population (59% in the study by Thane et al (44)) may have daily phylloquinone intakes below the current guideline of 1 [micro]g/kg body weight. There is now strong evidence that much higher dietary intakes are required for [gamma]-carboxylation of OC synthesized in bone than are required for [gamma]-carboxylation of the coagulation proteins synthesized in the liver. (4,27) A subclinical vitamin K deficiency might have a negative effect on bone health while causing no hemostatic consequences. (6,7) There is no current United Kingdom dietary recommendation for the role of vitamin K in the maintenance of bone health. Perhaps in recognition of the importance of this factor, the Institute of Medicine in the United States increased its recommendations for dietary vitamin K intake by ~50%, to 90 [micro]g/day for female subjects and 120 [micro]g/day for male subjects. (45) Sokoll and Sadowski (27) suggested that the normal adult population requires a daily intake between 100 and 400 [micro]g. A vitamin K intake that allows the vitamin K-dependent bone proteins to be in their maximal carboxylated state (6) is probably ~1000 [micro]g/day for healthy adults. (46)
Despite the loss of 60% to 70% of the daily dietary intake of phylloquinone through excretion, which emphasizes the need for a continuing daily supply to maintain tissue reserves, (47) vitamin K supplementation is not routine in United Kingdom CF clinics. Patients do receive routine vitamin A, D, and E supplements, however. There has been no consistently recommended level of vitamin K supplementation for patients with CF. The United States consensus statement of 1992 suggested mg weekly for infants, increasing thereafter to 5 mg weekly, (48) a dose that was subsequently shown to be inadequate. (49) More-recent recommendations from the United States, Europe, and the United Kingdom recognized that 5 mg 4 times per week is probably insufficient (50) and recommended a dose range between 1 mg/day and 10 mg/week (51) or 10 mg/day. (52) Authorities stressed the need for additional research before an optimal supplementary dose can be stated confidently.
This study is limited by the patients' suboptimal vitamin D levels. The CF unit protocol now demands a minimal value of >30 [micro]g/L. Only 13 study patients met this criterion. Twenty-six had levels between 20 and 30 [micro]g/L. High levels of Glu-OC may reflect vitamin D and vitamin K deficiency. (53) Vitamin D and calcium supplements can increase the percentage of Gla-OC among people with low BMD. (39) Uncoupling of the normal balance between bone formation and bone resorption has also been associated with a high vitamin A intake (54) and high serum retinol levels. (55) The median serum vitamin A level of the patient group was at the lower end of the normal range, with no patient showing a high value. Vitamin A imbalance was therefore unlikely to have contributed to any uncoupling of bone formation and resorption. Other potential influences on bone metabolism, including serum proinflammatory cytokine levels, respiratory exacerbation rates, calcium intake, corticosteroid use, activity levels, and general nutritional status, were not analyzed in this study.
Another limitation is the lack of comparative data for Glu-OC and Gla-OC levels from a control group of healthy children of the same age range. However, the levels of Glu-OC and Gla-OC in our population were comparable to those observed recently for non-vitamin K-supplemented Dutch CF patients, with the same direct assays. (9) The latter study showed that high-dose vitamin K supplementation ([greater than or equal to] 1 mg/day) decreased and increased the levels of Glu-OC and Gla-OC, respectively, to such a degree that the carboxylation status was improved even beyond that of the healthy control group. Another caveat is that we did not have a validated measure of total OC, although we did use BSAP and PICP and alternative, well-validated markers of bone formation. It was shown that the Takara Glu-OC MAb overestimates the measurement of the amino-terminal-mid-OC fragments and overestimates total OC levels, as represented by the sum of Glu-OC and Gla-OC measurements (. Shearer and C. Gundberg, unpublished data, 2004). As a consequence, the percentage of Glu-OC is also overestimated, compared with its measurement with binding assays. (8) However, studies among elderly subjects showed that the fractional decreases in the percentage of Glu-OC brought about with vitamin K supplementation were very similar whether measured with direct assays for Glu-OC and Gla-OC (as in the present study) or with indirect binding assays (. Shearer and C. Gundberg, unpublished data). These findings justify the use of the MAb-based assays for evaluating the carboxylation status of OC and its response to vitamin K supplementation. Future intervention studies might involve monitoring of vitamin K-induced changes in Glu-OC and Gla-OC levels, bone formation and resorption markers, and BMD.
CONCLUSIONS
This study showed that vitamin K deficiency is common among children with CF. Vitamin K levels showed a significant negative correlation with serum Glu-OC levels, which were associated with levels of bone turnover markers, suggesting that vitamin K has a role in maintaining the balance between bone formation and bone resorption. A cause-effect relationship between vitamin K deficiency and low bone mass is not proved. This study did not show any significant correlation between Glu-OC, Gla-OC, or vitamin K levels and any measure of BMD or BMC. Although PIVKA-II concentrations among children with CF have been normalized with vitamin K supplements, changes in the concentrations of bone turnover markers have yielded conflicting results. (9) Vitamin K supplementation among healthy adults may decrease Glu-OC levels without altering N-telopeptide or BSAP levels. (56) These conflicting results probably reflect the multifactorial pathogenesis of CF-related osteoporosis. It is difficult to isolate the discrete contribution to bone health of vitamin K deficiency. The supplemental dose of vitamin K necessary to achieve maximal [gamma]-carboxylation of OC among patients with CF and the effects of these supplements on bone turnover, BMD, and BMC remain to be determined in longitudinal and controlled studies.
Post a Comment
Note: only a member of this blog may post a comment.