Cystic fibrosis (CF) is a multisystem disease associated with defective bile salt absorption and pancreatic insufficiency, leading to steatorrhea and malabsorption of fat-soluble vitamins. With current management, life expectancy is well into adult life (1). However, this increase in life expectancy increases the risk of other diseases such as osteoporosis, which may directly or indirectly relate to the underlying disease state.
Routine supplementation of liposoluble vitamins is part of management, yet low serum levels of vitamins A and E, and to a lesser extent, of vitamin D (typically measured in its 25-OH form), are sometimes observed in CF (2). The reasons for this are not clear. We and others have found that vitamin levels do not always correlate with standard measurements of clinical status such as pulmonary function (forced expiratory volume in 1 second; FEV1) or growth and nutritional status (3,4), supporting the recommendation for routine biochemical assessment (5). Vitamin A and E levels depend on dietary intake, whereas vitamin D is related to both exposure to sunlight and dietary intake. Vitamin D derived from sunlight exposure can account for 80% of vitamin D levels (6). Furthermore, vitamin D also requires hydroxylation by both the liver (to form 25-OH vitamin D) and the kidney (to form 1,25-OH vitamin D) to fully exert its effect on bone metabolism.
It is currently recommended that vitamins A and E be monitored yearly (5). However, there are no existing recommendations for the monitoring of vitamin D. Given that vitamin D is fat soluble, as are vitamins A and E, we hypothesized that low levels of vitamins A and E may be associated with low 25-OH vitamin D, especially in climates where sunlight exposure is limited. If so, vitamin A and E levels could be used to identify patients at risk for a low 25-OH vitamin D level. This could eliminate the need for routine monitoring of 25-OH vitamin D, a more labor-intensive and costlier analysis that uses a radioactive tracer, so that testing could be performed in selected circumstances.
MATERIALS AND METHODS
We studied 40 consecutive charts of children with CF (21 girls) attending The Montreal Children's Hospital CF Clinic. These patients had been receiving vitamin and pancreatic enzyme therapy for at least 1 year, and measurements were obtained as part of routine clinical assessment. The reported values of vitamins A, E, and 25-OH vitamin D are those measured for the first time since the commencement of therapy. The diagnosis of CF was based on two positive sweat chloride test results (>60 mmol/l) and a compatible clinical history. The participants' mean age was 10.5 ± 3.90 (SD) years. None of the patients had evidence of biochemical liver disease (aspartate and alanine aminotransferases and γ-glutamyltransferase were within normal limits for our laboratory) and all were clinically stable.
Vitamin supplementation consisted of a multivitamin preparation containing both 5000 IU vitamin A (palmitate) and 400 IU D3, plus water-soluble vitamins. This was administered in a dose of one (<10 years of age) to two (>10 years of age) tablets per day. A water-miscible form of vitamin E (Aquasol E, Astra Merck, Wayne, PA, U.S.A.) was provided at a dose of 100 IU/day. At the time of blood testing, all patients also underwent testing for serum albumin, and serum cholesterol was assessed in 37. Thirty-three patients were old enough to undergo spirometric testing.
Height was measured by stadiometer with the patient in stocking feet. Weight was measured on an electronic balance with the patient lightly dressed. Weight was expressed as a percentage of ideal weight predicted for gender, age, and height, using the National Center for Health Statistics (NCHS) growth charts (5). Lung function, FEV1, was measured by expiratory spirometry (model 6200 Autobox D1; SensorMedics, Yorba Linda, CA, U.S.A.) and expressed as a percentage of predicted function (7).
Blood was collected after patients had fasted for at least 4 hours, with patients instructed to refrain from taking prescribed vitamins for the previous 24 hours. Vitamins A and E were measured by high-performance liquid chromatography (2), with a 25 cm × 4.6-mm, 5-μm column (LC-18; Supelcos, Sigma-Aldrich, Oakville, Ontario, Canada). Detection was at 292 nm (vitamin E) and 325 nm (vitamin A), with a photodiode array detector (Waters Chromatography, Ltd., Montreal, Canada). The interassay coefficient of variation was 5%. Serum 25-OH vitamin D was measured using a radioimmunoassay (Diasorin, Stillwater, MN, U.S.A.). The interassay coefficient of variation was 15%. Serum cholesterol and albumin were measured by a colorimetric technique using the laboratory's routine analyzer (Ektachem 750XRC, Ortho Diagnostics, Raritan, NJ, U.S.A.).
The number of patients with values below the lower limit of normal of percentage of predicted ideal weight (<85%), with moderate to severely impaired lung function (<60% FEV1), vitamin A (reference [ref.] range: 0–6 years, 0.70–1.50 μmol/l; 7–12 years, 0.90–1.70 μmol/l; and >12 years, 0.90–2.50 μmol/l) and vitamin E (ref. range: 1–6 years, 10–21 μmol/l; >6 years, 13–24 μmol/l) (8,9) were calculated. The vitamin E/cholesterol ratio was considered to be low if it was less than or equal to 4.8 mmol/mol (10). 25-OH vitamin D levels were considered to be low if they were less than 40 nmol/l, the lower limit for summer values in our laboratory. The summer lower limit was used, rather than the lower winter lower limit, because midrange values were thought to provide a margin of safety.
Data analysis was performed using a commercially available software package (Statistica for Windows, StatSoft, Inc., Tulsa, OK, U.S.A.). In each category, the number of normal and abnormal values was counted and respective percentages calculated. Descriptive statistics were used to examine the mean, minimum, and maximum values, and standard deviations. Summary frequency tables (χ2) were used to compare the proportion of patients with normal and low 25-OH vitamin D levels with the proportion of those with normal and abnormal measurements for vitamins A and E and vitamin E/cholesterol level. The relation between vitamin levels was assessed by the Pearson correlation coefficient. Comparisons between patients with low and normal levels of 25-OH vitamin D were made using two-tailed t-tests. P < 0.05 was considered significant.
Nine (22.5%) of 40 patients were malnourished (percentage of predicted ideal weight <85%) at the time of assessment, whereas 7 of 33 (21.2%) patients had moderate to severe lung disease (FEV1 <60%). Four of 40 (10%) patients had low levels of vitamin A. Three of 40 (7.5%) patients had low levels of vitamin E, whereas 16 of 37 (43.2%) patients had low cholesterol levels; 4 of 37 (10.8%) patients had a vitamin E/cholesterol level less than 4.8 mmol/mol. Six of 40 (15%) patients had low levels of albumin.
Four of forty (10%) patients had marginal or low 25-OH vitamin D (<40 nmol/l). The levels of 25-OH vitamin D varied between 25 and 150 nmol/l, with an average of 74.4 nmol/l. It was found that all four patients with low 25-OH vitamin D were older, with no child less than 12 years of age having a 25-OH vitamin D less than 40 nmol/l. Of those 36 patients with normal 25-OH vitamin D levels, 10 (27.8%) were more and 26 (72.2%) were less than 12 years of age. The mean age of children with normal 25-OH vitamin D levels was 10.0 years, whereas it was 15.0 years for those with low levels (P < 0.01;Table 1). The group with low levels of 25-OH vitamin D and the group with normal levels did not differ in percentage of predicted ideal weight (P = 0.88), lung function (percentage of predicted FEV1;P = 0.36), or vitamin A levels (P = 0.71).
Children with low levels of 25-OH vitamin D had lower vitamin E and vitamin E/cholesterol levels when compared with those with normal 25-OH vitamin D levels (Table 1). Frequency distributions of the various vitamin levels are shown in Table 2. Low vitamin E level had a positive predictive value for a low 25-OH vitamin D level of 66.7%, whereas a normal vitamin E level had a negative predictive value of 94.6%. The corresponding values for the low vitamin E/cholesterol level were 50% and 93.9%; for vitamin A, the values were 0% and 88.9%.
25-OH vitamin D levels correlated significantly with vitamin E/cholesterol levels (r = 0.41, P < 0.01, n = 37), and weakly with vitamin E levels (r = 0.28, P < 0.08, n = 40), but not with vitamin A levels (r = −0.14, P > 0.1, n = 40;Figs. 1 and 2). Vitamin A levels correlated significantly with both vitamin E (r = 0.38, P < 0.01) and vitamin E/cholesterol levels (r = 0.35, P < 0.03). There was no significant correlation between 25-OH vitamin D levels and lung function (percentage of predicted FEV1:r = 0.07, P > 0.1), or nutritional status (percentage of predicted ideal body weight:r = −.02, P > 0.1), but there was a weak inverse correlation with age (r = −0.29, P < 0.07).
Our results indicate that children less than 12 years of age, or those with normal vitamin E levels, are unlikely to have low 25-OH vitamin D levels. These results are consistent with those of previous studies, in that subnormal levels of 25-OH vitamin D in children with CF are relatively rare, and neither pulmonary disease nor nutritional status correlates with levels of 25-OH vitamin D (4).
We examined the relationship between vitamin levels in two ways. The first was to consider the values as dichotomous, either above the lower limit of normal or below. Although in fact these values are continuous, clinically they are treated in comparison with a reference range. Viewed in this manner, vitamin E levels above the lower limit were strongly predictive of values in the normal range, and values below the normal range were predictive of low 25-OH vitamin D levels. However, in contrast to previous results (4), age seemed to play a factor in determining those with low levels of 25-OH vitamin D in this study. Looked at as continuous variables, 25-OH vitamin D levels correlated with vitamin E/cholesterol and weakly with vitamin E levels, but not with vitamin A levels, even though vitamin A levels correlated with vitamin E levels.
Vitamins A and E are derived primarily from absorption through the gastrointestinal tract, whereas vitamin D is derived principally from synthesis in the skin in temperate climates (11,12). However, low levels of vitamin A may not be the result of malabsorption alone but also its mobilization from liver stores. Vitamin A is mobilized from liver stores and transported in plasma as retinol bound to retinol-binding protein. This complex interacts strongly with transthyretin (prealbumin) and normally circulates as a 1:1 molar retinol-binding protein-transthyretin complex (13). In uncomplicated vitamin A deficiency, serum levels are maintained at the expense of liver stores. In patients with CF, however, low serum levels may be accompanied by high levels in the liver (14). Low plasma levels of both vitamin A and retinol-binding protein have been readily demonstrated in patients with CF (2). It has been speculated that there may be a defect in the liver's synthesis of retinol-binding protein or a problem with the release of this protein from the liver (14). Besides alterations in serum handling, it has been shown that vitamin A loss in stool of children with CF is prominent even while they are receiving pancreatic supplementation (15). In addition, pancreatic-sufficient patients with CF may have low vitamin A levels with normal vitamin E levels, perhaps related to low circulating levels of retinol-binding protein and prealbumin (16). This complex metabolism of vitamin A is perhaps the reason vitamin A deficiency cannot be used to predict vitamin D status in children. However, it does not explain the correlation between vitamin A and E levels; rather, it explains the absence of correlation between vitamin A and 25-OH vitamin D levels, even though they are provided in the form of a combined vitamin preparation.
In contrast, vitamin E is relatively poorly absorbed, even in healthy individuals. It is transported to peripheral tissues by low-density lipoprotein (LDL), where it is concentrated in the membranes of cells and organelles. The major storage site is in adipose tissue (17). Patients with CF have lowered fat masses and thus diminished storage ability, making them more dependent on intestinal absorption, so that regular intake of vitamin E is typically recommended, in a water-miscible form. Future work is necessary to optimize dosage and levels of vitamin E to provide maximal protection against oxidant stress (10,18).
Vitamin E status in patients with CF also depends on the degree of pancreatic sufficiency (19) and compliance. Although we did not intend that vitamin E serve as a marker of compliance with vitamin supplementation, in retrospect we found that it was generally patients known to be noncompliant (irregular or nonintake of prescribed vitamins) who had subnormal levels of vitamin E. We did not formally question the patients about their adherence to vitamin prescriptions. However, patients who rarely take their vitamin supplements would be at risk for vitamin D deficiency, especially if they had limited exposure to sunlight. A study from Ireland found that three quarters of adult patients with CF had subnormal 25-OH vitamin D levels (at least during the winter months) (20). Perhaps the supplementation of milk with vitamin D in North America lessens the risk of decreased sunlight exposure, even in the CF population. If the issue of vitamin levels were related solely to compliance, then we would have expected all three vitamin levels to correlate, which was not the case. This again suggests that metabolism of each vitamin and its environmental sources (nutritional, sunlight exposure) also exerts an influence on circulating levels.
The development of osteoporosis in those with CF is of increasing importance, and several factors are known to play a part. Deficiency of 25-OH vitamin D is a risk factor (21), but osteoporosis may develop in patients with CF despite normal 25-OH vitamin levels. Nutritional status influences the weight stress on bones, so that lighter individuals have a lower bone mass (21,22). The degree of pulmonary impairment also correlates with bone mineralization (21). This may be because of the correlation between nutritional and pulmonary status in patients with CF (23). However, it may also be due to the effect of the chronic inflammatory state that characterizes CF lung disease (24) and its influence on bone formation and resorption (25,26). The level of physical activity can also influence bone mineral density in children (27). Of note, we have shown that in those patients with CF with significant airflow limitation (FEV1 <75% of predicted function), the percentage of time spent in vigorous activity was related to their nutritional status (28). This again demonstrates the interplay between factors influencing bone mineral density. Due to these multiple factors, it probably would be most useful to routinely monitor bone mineral density directly, using dual energy x-ray absorptiometry (DEXA).
A recent study in adult patients with CF showed a very high incidence of 25-OH vitamin D deficiency, despite being given vitamin supplementation, and those who were clearly deficient had markedly low bone mineral density (29). These patients were under consideration for lung transplantation and therefore had both poor lung function and poor nutritional status. No correlation between levels of 25-OH vitamin D and bone mineral density was observed in patients 17 to 42 years of age, despite 5 of 22 having below normal vitamin D levels (30). In this study, nutritional status and disease severity related to bone mineral density. However, the use of linear regression analysis (31) may not be an optimal way of determining the relation between vitamin D status and bone mineral density (21), because vitamin levels would cease to have a limiting effect once adequate levels are achieved.
We found relatively infrequent vitamin deficiencies in our patients. This may not be applicable to other CF populations, especially in regions where milk is not supplemented with vitamin D, sunlight is limited, or vitamins are not routine prescribed. We also did not question our patients about compliance, nor did we assess the degree of steatorrhea. These are less important issues when levels are normal but are often the most likely explanations for low levels.
In conclusion, 25-OH vitamin D levels are rarely low in North American children with CF who receive multivitamin supplementation and routine supplementation of milk with vitamin D. In particular, children less than 12 years of age and older children with normal vitamin E levels are unlikely to have low 25-OH vitamin D levels, and the measurement can therefore probably be omitted in them. In contrast, those children with low vitamin E levels are at an increased risk for low 25-OH vitamin D levels and thus may warrant monitoring.
The authors thank Marino A. Perlas for technical assistance.
1. Davis PB, Drumm M, Konstan MW. Cystic fibrosis. Am J Respir Crit Care Med 1996; 154:1229–56.
2. Peters SA, Rolles CJ. Vitamin therapy in cystic fibrosis—a review and rationale (review). J Clin Pharm Ther 1993; 18:33–8.
3. Grey VL, Lands LC, Grenier C, Drury D. Vitamin levels and clinical status in cystic fibrosis [abstract]. Clin Biochem 1997; 30:370.
4. Henderson RC, Lester G. Vitamin D levels in children with cystic fibrosis. South Med J 1997; 90:378–83.
5. Ramsey BW, Farrell PM, Pencharz P. Nutritional assessment and management in cystic fibrosis: A consensus report. Am J Clin Nutr 1992; 55:108–16.
6. Rayner RJ. Fat-soluble vitamins in cystic fibrosis. Proc Nutr Soc 1992; 51:245–50.
7. Polgar G, Promadhat V. Pulmonary function testing in children.
Philadelphia: WB Saunders Co., 1971.
8. Lockitch G, Halstead AC, Wadsworth L, Quigley G, Reston L, Jacobson B. Age-and sex-specific pediatric reference intervals and correlations for zinc, copper, selenium, iron, vitamins A and E, and related proteins. Clin Chem 1988; 34:1625–8.
9. Lockitch G, Halstead AC, Albersheim S, MacCullum C, Quigley G. Age-and sex-specific pediatric reference intervals for biochemistry analytes as measured with the Ektachem-700 analyser. Clin Chem 1988; 34:1622–5.
10. James DR, Alfaham M, Goodchild MC. Increased susceptibility to peroxide-induced haemolysis with normal vitamin E concentrations in cystic fibrosis. Clin Chim Acta 1991; 204:279–90.
11. Hubbard VS, Farrell PM, di Sant'Agnese PA. 25-Hydroxycholecalciferol levels in patients with cystic fibrosis. J Pediatr 1979; 94:84–6.
12. Reiter EO, Brugman SM, Pike JW, Pitt M, Dokoh S, Haussler MR, et al. Vitamin D metabolites in adolescents and young adults with cystic fibrosis: Effects of sun and season. J Pediatr 1985; 106:21–6.
13. Goodman DS. Vitamin A and retinoids in health and disease. N Engl J Med 1984; 310:1023–31.
14. Durie PR, Pencharz PB. A rational approach to the nutritional care of patients with cystic fibrosis. J Roy Soc Med 1987; 80:25–9.
15. Ahmed F, Ellis J, Murphy J, Wootton S, Jackson AA. Excessive faecal losses of vitamin A (retinol) in cystic fibrosis. Arch Dis Child 1990; 65:589–93.
16. Lancellotti L, D'Orazio C, Mastella G, Mazzi G, Lippi U. Deficiency of vitamins E and A in cystic fibrosis is independent of pancreatic function and current enzyme and vitamin supplementation. Eur J Pediatr 1996; 155:281–5.
17. Oslo RE. Disorders of fat soluble vitamins A, D, E, K. In: Suskind RM, Lewinter-Suskind L, eds. Textbook of pediatric nutrition. New York: Raven Press, 1997: 49–72.
18. Winklhofer-Roob BM, Ziouzenkova O, Puhl H, Ellemunter H, Greiner P, Muller G, et al. Impaired resistance to oxidation of low density lipoprotein in cystic fibrosis: Improvement during vitamin E supplementation. Free Radic Biol Med 1995; 19:725–33.
19. Kalnins D, Corey M, Durie PR, Ellis L, Ellis G. Do serum vitamin E levels correlate with 72-hour fecal fat in assessing pancreatic function at time of CF diagnosis (abstract). Pediatr Pulmonol 1995; 12(Suppl):A308.
20. Hanly JG, McKenna MJ, Quigley C, Freaney R, Muldowney FP, Fitzgerald MX. Hypovitaminosis D and response to supplementation in older patients with cystic fibrosis. Q J Med 1985; 56:377–85.
21. Henderson RC, Madsen CD. Bone density in children and adolescents with cystic fibrosis. J Pediatr 1996; 128:28–34.
22. Gibbens DT, Gilsanz V, Boechat MI, Dufer D, Carlson ME, Wang CI. Osteoporosis
in cystic fibrosis. J Pediatr 1988; 113:295–300.
23. Vaisman N, Pencharz PB, Corey M, Canny GJ, Hahn E. Energy expenditure of patients with cystic fibrosis. J Pediatr 1987; 111:496–500.
24. Cantin A. Cystic fibrosis lung inflammation: Early, sustained, and severe (editorial;comment). Am J Respir Crit Care Med 1995; 151:939–41.
25. Papanicolaou DA, Wilder RL, Manolagas SC, Chrousos GP. The pathophysiologic roles of interleukin-6 in human disease. Ann Intern Med 1998; 128:127–37.
26. Hyams JS, Wyzga N, Kreutzer DL, Justinich CJ, Gronowicz GA. Alterations in bone metabolism in children with inflammatory bowel disease: An in vitro study. J Pediatr Gastroenterol Nutr 1997; 24:289–95.
27. Kroger H, Kotaniemi A, Vainio P, Alhava E. Bone densitometry of the spine and femur in children by dual-energy x-ray absorptiometry. Bone Miner 1992; 17:429–39.
28. Boucher GP, Lands LC, Hay JA, Hornby L. Activity levels and the relationship to lung function and nutritional status in children with cystic fibrosis. Am J Phys Med Rehabil 1997; 76:311–5.
29. Donovan DS, Papadopoulos A, Staron RB, Addesso V, Schulman L, McGregor C, Cosman F, et al. Bone mass and vitamin D deficiency in adults with advanced cystic fibrosis lung disease. Am J Respir Crit Care Med 1998; 157:1892–9.
30. Grey AB, Ames RW, Matthews RD, Reid IR. Bone mineral density and body composition in adult patients with cystic fibrosis. Thorax 1993; 48:589–93.
31. Aris RM, Renner JB, Winders AD, Buell HE, Riggs DB, Lester GE, et al. Increased rate of fractures and severe kyphosis: Sequelae of living into adulthood with cystic fibrosis. Ann Intern Med 1998; 128:186–93.