Vitamin A levels in patients with CF are influenced by the inflammatory response

Journal of Cystic Fibrosis 3 (2004) 143 – 149 www.elsevier.com/locate/jcf

Vitamin A levels in patients with CF are influenced by $ the inflammatory response Ristan M. Greer a,*, Helen M. Buntain a,b, Peter J. Lewindon c, Claire E. Wainwright b, Julia M. Potter e, Joseph C. Wong f, Paul W. Francis b, Jennifer A. Batch d, Scott C. Bell g,h a

Department of Paediatrics and Child Health, University of Queensland, Brisbane, Australia b Respiratory Medicine, Royal Children’s Hospital, Brisbane, Australia c Gastroenterology and Hepatology, Royal Children’s Hospital, Brisbane, Australia d Diabetes and Endocrinology, Royal Children’s Hospital, Brisbane, Australia e Queensland Health Pathology Services, Brisbane, Australia f Nuclear Medicine, Royal Brisbane Hospital, Brisbane, Australia g Thoracic Medicine, The Prince Charles Hospital, Brisbane, Australia h Department of Medicine, University of Queensland, Brisbane, Australia Received 3 October 2003; accepted 27 April 2004 Available online 27 July 2004

Abstract Background: Serum vitamin A, normally depressed in inflammatory conditions, is frequently low in people with CF. Vitamin A is important in respiratory epithelial regeneration and repair. We hypothesised that serum vitamin A would be associated with inflammation and disease severity. Methods: Serum vitamin A (as retinol), C-reactive protein (CRP), vitamin E, 25-hydroxy vitamin D (25OHD), 1,25dihydroxy vitamin D (1,25(OH)2D), weight, and lumbar spine bone mineral density (LSBMD) were measured in 138 subjects with CF (5 – 56 years) and 138 control subjects (5 – 48 years). FEV1, presence of CF liver disease (CFLD) and hospital admissions were recorded in those with CF. Results: Serum vitamin A level was lower in CF subjects than in controls (mean, 95% CI: 1.29, 1.0 – 1.37 vs. 1.80, 1.7 – 1.87 Amol/l, p < 0.0001), and inversely correlated with CRP (rs = 0.37, p < 0.0001). CF subjects with low vitamin A (45%) level had poorer FEV1, weight z-score, LSBMD z-score, and higher CRP compared with those with normal levels. In the CF group CRP, vitamin E, 1,25(OH)2D, presence of CFLD, admissions, and age were associated with vitamin A level. Conclusions: Serum vitamin A is negatively associated with CRP in subjects with CF, consistent with normal population studies. It is important to distinguish between low serum vitamin A associated with the inflammatory response and that due to poor nutritional stores. The role of vitamin A in CF warrants further study, in the contexts both of chronic recurrent inflammatory disease and acute pulmonary exacerbation. D 2004 European Cystic Fibrosis Society. Published by Elsevier B.V. All rights reserved. Keywords: Cystic fibrosis; C-reactive protein; Inflammation; Retinol

1. Introduction People with cystic fibrosis (CF) are at risk for deficiency of the fat soluble vitamins (A, D, E and K) [2]. Serum

$ This work was presented at the 26th European Cystic Fibrosis Conference, Belfast, June 2003 [1]. * Corresponding author. Department of Paediatrics and Child Health, University of Queensland, Royal Children’s Hospital, Herston, Queensland 4029, Australia. Tel.: +61-7-3365-5338; fax: +61-7-3365-5455. E-mail address: [email protected] (R.M. Greer).

vitamin A levels are often suboptimal [3], and low levels have been described in infants newly diagnosed by neonatal screening [4,5]. Vitamin A metabolism is altered during the acute phase response, where serum levels of vitamin A and retinol binding protein fall transiently due to decreased hepatic release, concurrent with increases in acute phase proteins such as C-reactive protein (CRP) [6,7]. ‘Vitamin A’, the generic term for retinoids, includes retinol (alcohol form), retinal (aldehyde form) and retinoic acid (acid form). Dietary carotenoids including h-carotene are included in the vitamin A family, as they are converted

1569-1993/$ - see front matter D 2004 European Cystic Fibrosis Society. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.jcf.2004.04.003

144

R.M. Greer et al. / Journal of Cystic Fibrosis 3 (2004) 143–149

to retinol in the small intestinal mucosa, 6 Ag h-carotene providing 1 Ag retinol [8]. Biologically, vitamin A is important for vision, epithelial differentiation and function, apoptosis, the immune system, resistance to infectious disease, and its’ precursor h-carotene has antioxidant properties. Acute inflammation causes a reduction in serum vitamin A, ascribed to a reduction in the hepatic synthesis of retinol binding protein, and reduction in secretion of the retinol – retinol binding protein complex [6,9,10]. Conversely, animal studies have shown a protective effect of both retinol (vitamin A) and hcarotene supplementation against the inflammatory response [11,12]. Vitamin A deficiency causes night blindness and xerophthalmia, and is the leading cause of childhood blindness in developing countries [13]. Night blindness and xerophthalmia have been described in CF patients [14,15]. Vitamin A appears to have a role in protection of the respiratory tract and supplementation has reduced morbidity and mortality in infants from developing countries [16]. CF lung disease is characterised by excessive inflammatory response to infection, which is present even in presymptomatic infants [17,18]. Inflammation is accompanied by neutrophilia with production of elastase and proinflammatory cytokines such as interleukins 6 and 8 [19]. We have observed that serum vitamin A levels, below the normal laboratory range in patients with severe disease, returned to normal post lung transplant independent of oral supplementation or pancreatic sufficiency [20]. We hypothesised that serum vitamin A is associated with inflammation and disease severity, and that vitamin A levels would decrease as disease severity increases with age.

2. Materials and methods 2.1. Subjects Data from the Royal Children’s Hospital paediatric and The Prince Charles Hospital adult CF clinic subjects (n = 138 CF (72 M) and 138 control (61 M), aged 5 – 56 years), in whom serum vitamin A level had been measured, was used. Subject recruitment, and methods for vitamin D, anthropometric indices and bone mineral density (BMD) have been described previously [21]. Briefly, patients with CF attending the Royal Children’s Hospital and The Prince Charles Hospital Adult CF Clinic, Brisbane, were invited to participate. Paediatric control subjects were recruited by invitation from local regional schools. Adult control subjects were recruited from family and friends of the investigators and subjects with CF. Most CF subjects were studied as outpatients, except where rural or regional subjects had been admitted for routine review or pulmonary exacerbation, in which case investigations (blood samples and bone density measurements) were undertaken after the completion of any therapy, at least 14 days post-admission, prior to discharge. A diagnosis of CF was confirmed with a

positive sweat test. Blood samples for measurement of serum vitamin A were protected from light. Bone density at the lumbar spine was measured with a Lunar DPX-L densitometer (Madison, Wisconsin). Measurements for serum vitamin A (vitamin A), vitamin E, 25-hydroxy vitamin D (25OHD), 1,25 di-hydroxy-vitamin D (1,25(OH)2D), CRP, anthropometric indices, BMD and lung function (% predicted forced expiratory volume in 1 s, FEV1) at the time of blood collection were available. Serum vitamin A (measured as retinol) and vitamin E were measured by an in house chromatography method used by Queensland Health Pathology Services. Using this method, plasma is treated with alcohol to release protein-bound retinol, which is taken up into ethanol and chromatographed on HPLC (Waters Alliance HPLC Integrated System). Peaks are monitored by UV detection as 330 nm (Beckman DU 640 Spectrophotometer). The coefficient of variation for measurement of serum vitamin A is 2.6%. CRP was measured using the Beckman IMMAGE Immunochemistry system. Serum CRP binds to particle-bound anti-CRP antibody, forming an antigen – antibody complex. The IMMAGE system measures the rate of increase in light scattered from the antigen –antibody particle complexes. This test has a sensitivity (lowest detectable concentration) of 0.1 mg/dl and a coefficient of variation of 6.5%. CRP was defined as elevated if >5 mg/l. Pseudomonas aeruginosa (PSA) infection, number of hospital admissions, pulmonary exacerbations and number of days in hospital for the previous 12 months were recorded. The presence of CF related liver disease (CFLD), defined as evidence of cirrhosis on liver biopsy or portal hypertension on ultrasound, was recorded. Subjects with CF were routinely prescribed 1650– 5000 IU vitamin A (or equivalent) daily as part of normal clinical care. Four patients were on supplemental doses of up to 50,000 units twice weekly. Serum vitamin A levels were defined as low if they were under the lower limit of the normal range for our laboratory (1.0 – 2.3 Amol/l for subjects under 16 years, and 1.6– 2.3 Amol/l for subjects 16 years or over). The study was approved by the ethics committees of the Royal Children’s Hospital and The Prince Charles Hospital. After provision of verbal and written information about the study, written consent was obtained from adults, parents or carers of children, and from those children old enough to give consent. 2.2. Statistics Variables were described as mean F standard deviation or age-adjusted means with 95% confidence intervals (CI) where levels varied with age. Non-normally distributed variables (CRP) were described as median with interquartile range. Groups were compared using t-tests or analysis of covariance to adjust for age, with control status (control or CF) and gender considered as main effects. Tukey’s adjustment was used for post hoc multiple comparisons.

R.M. Greer et al. / Journal of Cystic Fibrosis 3 (2004) 143–149

145

Table 1 Characteristics of control subjects and subjects with CF by age group Children

M/F Age (years) Weight z-score Height z-score FEV1 %

Adolescents

Control

CF

10/15 8.7 F 1.5 0.35 F 1.2 0.30 F 1.1

18/21 8.5 F 1.8 0.03 F 1.0 0.24 F 0.9 75.4 F 16.9

Adults

p

Control

CF

0.58 0.24 0.04

30/35 14.0 F 2.0 0.24 F 0.8 0.60 F 1.0

26/19 14.3 F 2.2 0.82 F 0.9 0.89 F 0.9 69.0 F 25.7

Correlations were examined using Spearman’s rank correlation coefficient (rs) with adjustment (partial correlation) for age, in order to include CRP in the correlation matrix. Wilcoxon’s test was used to compare CRP levels. Because physiology is different at various stages of growth, subjects were divided in some instances into age groups of children < 11 years, adolescents 11– 18 years and adults >18 years. Predictors of serum vitamin A were examined using multiple regression. Two multiple regression models were examined, the first including both control and CF subjects, and the second restricted to CF subjects to allow variables specific to the CF group to be included in the model. In the first multiple regression model control status, weight zscore, vitamin E, CRP, 25OHD, 1,25(OH)2D, age, and interaction terms for age and control status and CRP and control status were entered as independent variables. In the second model, restricted to CF subjects, control status was omitted and FEV1, PSA infection, presence of CFLD, number of admission, hospitalisation and exacerbation episodes included as independent variables, otherwise the same predictor variables were used in both models. To further characterise the evolution of vitamin A levels with age and stages of growth, we also used multiple regression with backward selection to investigate the interactions between control status and vitamin A level in children, adolescents and adults separately. Odds ratios were used to estimate the risk of low serum vitamin A in those with high CRP. SAS version 8.0 (SAS Institute, Cary, NC) was used for all analyses.

p

Control

CF

p

0.46 < 0.0001 < 0.0001

21/27 27.0 F 7.5 0.06 F 0.5 0.03 F 0.4

28/26 27.3 F 8.0 0.34 F 0.9 0.24 F 1.0 54.1 F 20.5

0.81 0.21 0.01 0.16

25OHD is the precursor for 1,25(OH)2D. Weight z-score reflects nutritional status. FEV1, number of admissions, hospitalisations and exacerbations are all reflectors of clinical status and disease severity, while CRP is a marker for inflammation. CFLD may affect dynamics of vitamin A because of potentially altered secretion of retinol [10].

3. Results 3.1. Vitamin A Study subjects characteristics are described in Table 1. Vitamin A was lower in CF subjects overall compared with controls (1.29, 1.0 – 1.37 vs. 1.80, 1.7 – 1.87 Amol/l, p < 0.0001), increasing with age in both groups ( p < 0.0001 for relationship with age). There was no difference in vitamin A level between genders in either the control or the CF groups (Table 2). When analysed by age group, vitamin A levels were lower in adolescents and adults with CF compared with controls, but not significantly different in children. In children and adolescents, there was an interaction between control status and age such that vitamin A level increased with age in control subjects to a greater degree than in CF subjects (Fig. 1).

2.3. Selection of predictive variables for vitamin A Vitamin E, like vitamin A, is a fat-soluble anti-oxidant and might be expected to be associated with vitamin A in these patients with pancreatic and intestinal malabsorption. 1,25(OH)2D interacts with retinoic acid metabolism [22]. Table 2 Mean, 95% confidence interval, for vitamin A (Amol/l), in male and female CF and control subjects

Male Female p

Control subjects: mean, CI

CF subjects: mean, CI

p

n = 61, 1.85, 1.72 – 2.0 n = 77, 1.76, 1.62 – 1.90 0.30

n = 72, 1.34, 1.22 – 1.45 n = 66, 1.24, 1.11 – 1.37 0.30

< 0.0001 < 0.0001

Fig. 1. Mean (95% CI) serum vitamin A (Amol/l) in control subjects and subjects with CF. Serum vitamin A was not significantly different between control and CF children (1.45, 1.25 – 1.64, n = 25 vs. 1.11, 0.95 – 1.27, n = 39, p = 0.08) but higher in control adolescents (1.73, 1.61 – 1.85, n = 65 vs. 1.19, 1.05 – 1.34, n = 45, p < 0.0001) and adults (2.08, 1.94 – 2.22, n = 48 vs. 1.50, 1.37 – 1.63, n = 54, p < 0.0001) compared with the CF group.

146

R.M. Greer et al. / Journal of Cystic Fibrosis 3 (2004) 143–149

3.2. Prevalence of low serum vitamin A The prevalence of low serum vitamin A was higher in the CF group compared with controls ( p < 0.0001). Forty-five percent (62/138, 32 M) of CF subjects had low vitamin A and 55.1% (76/138, 34 M) had normal levels, compared with 11.6% (16/138, 12 M) of control subjects with low levels and 88.4% (122/138, 65 M) with normal levels. In the CF group, 30.8% (12/39) children, 44.5% (20/45) adolescents and 55.6% (30/54) adults had a low level of vitamin A. In the control group, 16.0% (4/25) children, 3.1% (2/65) adolescents and 20.8% (10/48) adults had low levels. The proportion of subjects with low vitamin A was different according to age group in the control subjects ( p = 0.01), but similar in the CF group ( p = 0.06). In children, prevalence of low vitamin A was similar in the control and CF groups ( p = 0.18) but prevalence was increased in adolescents ( p < 0.0001) and adults ( p = 0.0003) with CF compared with controls. Ten subjects had CFLD, 3 children (2 M), 3 adolescents (3 M), and 4 adults (2 M). Serum vitamin A level was lower in the group with CFLD (0.82, 0.48 – 1.17, vs. 1.57, 1.50– 1.64 Amol/l, p = 0.003). A greater proportion of subjects with CFLD had low vitamin A level (10/14 (71.4%) compared with 52/124 (42%) without CFLD, p = 0.04). There was no difference in vitamin A level between pancreatic insufficient (n = 130, 70 M) and sufficient subjects with CF (n = 8, 2 M) (1.28 F 0.5 vs. 1.36 F 0.8 Amol/l, p = 0.80). There was no difference in vitamin A level between the 4 subjects with CF who were on high dose vitamin A supplementation compared with those on routine supplementation, nor any difference between the 11 patients (9 children, 2 adults) free of PSA and the other patients. In CF subjects, FEV1 (70.6 F 21.8 vs. 57.9 F 23.1 %, p = 0.001), weight z-score ( 0.12 F 0.98 vs. 0.76 F 0.94, p = 0.0002), and lumbar spine BMD z-score ( 0.50 F 1.24 vs. 1.08 F 1.31, p = 0.009) were lower in patients with low vitamin A. There was no difference in height z-score between those with low vitamin A and those with vitamin A in the normal range ( 0.35 F 0.91 vs. 0.55 F 1.13, p = 0.27). CRP was higher (median = 7.0, IQR = 2.0 – 14.0 vs. median = 1.0 IQR = 1.0 – 3.0 mg/l, p < 0.0001) in CF subjects with low vitamin A. There were no gender differences in FEV1, weight z-score or CRP in those with low vitamin A vs. those with levels in the normal range. Lumbar spine BMD z-score was lower in male subjects with low vitamin A (n = 30) vs. those with normal levels (n = 42) ( 1.53 F 1.44 vs. 0.58 F 1.42, p = 0.009), but no different in females with low (n = 32) vs. normal (n = 34) vitamin A levels ( 0.66 F 1.03 vs. 0.43 F 0.99, p = 0.88). Thus the difference in bone mineral density was attributable mainly to the male group. Lumbar spine BMD z-score was similar in males and females with normal vitamin A levels ( p = 0.96), but lower in males than in females in those with low levels ( p = 0.03). CRP was available in 133 subjects with CF, high in 51, and in 136 control subjects, high in 17. CRP was correlated

Table 3 Multiple regression for predictors of vitamin A in control subjects and subjects with CF Predictive independent variables

Regression coefficient F S.E.

(a) Model 1—includes control and CF subjects Control status (control, 0.42 F 0.07 CF coded as 0,1) Vitamin E (Amol/l) 0.02 F 0.004 CRP (mg/l)  control 0.02 F 0.005 status interaction 0.002 F 0.0007 1,25(OH)2D (pmol/l) Age (years) 0.02 F 0.004 (b) Model 2—restricted to CF subjects Vitamin E (Amol/l) CRP (mg/l) 1,25(OH)2D (pg/l) CF related liver disease Number of hospital admissions Age (years)

0.02 F 0.004 0.02 F 0.004 0.003 F 0.0009 0.43 F 0.11 0.04 F 0.02 0.02 F 0.004

p

< 0.0001 0.0002 0.0006 0.0005 < 0.0001

< 0.0001 < 0.0001 0.004 0.0001 0.03 < 0.0001

with age in the CF group (rs = 0.5, p < 0.0001), but not in controls (rs = 0.06, p = 0.45). CRP was higher in CF subjects compared with control subjects (median 3.0, IQR 1.0– 10.0, n = 133, vs. median 2.0, IQR 1.0 – 3.0 mg/l, n = 136, p = 0.04). CRP was negatively correlated with vitamin A in CF subjects (rs = 0.37, p < 0.0001), but not in control subjects (rs = 0.01, p = 0.9), and correlated in CF subjects with FEV1 (rs = 0.42, p < 0.0001). 3.3. Predictors of vitamin A level In the first multiple regression model, control status, vitamin E, 1,25(OH)2D, age, and the interaction between CRP and control status were predictive of vitamin A level, the interaction term indicating that CRP was predictive in CF but not in control subjects (Table 3a). Weight z-score and 25OHD vitamin D were not significant predictors. These predictors explained 36% of the variation in vitamin A (r2 = 0.36). In the second multiple regression model restricted to the CF group, CRP, vitamin E, and 1,25(OH)2D, presence of CFLD, number of admissions, and age, but not FEV1, weight z-score, number of hospitalisations or exacerbations, were predictive (Table 3b). Significant predictors explained 49% of the variation in vitamin A (r2 = 0.49). High CRP was a risk factor for low serum vitamin A in the CF group (OR 6.39, CI 2.93 – 13.90), where 37 of those with high CRP had low vitamin A. None of the control subjects with high CRP had low vitamin A.

4. Discussion This study demonstrates that serum vitamin A levels are inversely proportional to CRP in subjects with CF. This is consistent with reports from large epidemiological studies

R.M. Greer et al. / Journal of Cystic Fibrosis 3 (2004) 143–149

based on the general population [23]. The implication is that decreased serum retinol associated with inflammation is not caused by nutritional vitamin A deficiency. Indeed, in studies aimed at measuring prevalence of nutritional vitamin A deficiency, presence of inflammation is a confounder. In non-CF populations, serum vitamin A decreases transiently, proportionate to the rise in CRP, as part of the normal acute phase response to inflammation [6,24]. The inverse relationship between serum retinol and inflammation is notably well documented in Stephensen and Gildengorin’s analysis of data from more than 22,000 subjects in large third National Health and Nutrition Examination Survey (NHANES III) [23], as well as in smaller studies of specific inflammatory episodes [24]. The mechanism for these changes is not at present entirely understood, but several factors have been elucidated. Hepatic vitamin A release is decreased concurrent with impaired synthesis or release of retinol binding protein and transthyretin from the liver [6,7,10,24]. Oxidative stress and free radical production associated with inflammation may increase metabolic requirement for vitamin A [25], and urinary retinol loss during severe infections has been associated with decreased serum retinol [26]. Cystic fibrosis is a case of inflammatory disease where the insult is repeated, and associated with progressive tissue damage, complicated by liver disease in around 10– 15% of patients. A re-evaluation of the role of vitamin A in CF is required. Nutritional deficiency must be differentiated from low serum retinol associated with the inflammatory response. Nutritional vitamin A deficiency in CF subjects has been clearly documented in the literature. Low serum retinol levels, which recovered with supplementation and enzyme replacement, have been reported in infants with CF diagnosed by neonatal screening [4,27]. Night blindness and xerophthalmia have been reported [14,15], although these complications appear rare. Their association with serum vitamin A level is likely to be multifactorial, and prospective studies are needed. In recent years, enhanced understanding of nutritional issues in CF, and the widespread recommendation of routine daily vitamin A supplementation [2] perhaps make the occurrence of true vitamin A deficiency in CF less likely. However issues of imperfect medication adherence [28], and the prolonged nature of the inflammatory process in CF, suggest that some subjects may be at risk of suboptimal stores. Lindblad et al. [29] documented decreasing concentration of vitamin A in the liver with age. It is likely that in clinical practice there is an overlap in these two scenarios, with some patients having nutritional deficiency and low liver stores, as well as low serum levels associated with inflammation, while other patients have adequate liver stores with low serum levels associated with inflammation alone, but not nutritional deficiency. This is exemplified in the current study, where 61% of CF patients with serum vitamin A level below the normal range also had elevated CRP, however the remaining 39% of CF subjects with low vitamin A had normal CRP. Despite the possibility that adoption of an arbitrary cutoff point in

147

children and adults may engender some misclassification, this observation suggests that a number of subjects had low serum vitamin A not associated with the acute inflammatory response. The overlap between nutritional and inflammatory changes in serum vitamin A level is important, because nutritional vitamin A deficiency (possibly compounded by increased requirement) is theoretically treatable by increased supplementation or improved attention to adherence. Also, pulmonary exacerbation in patients with prolonged inflammatory disease, and pre-existing low or marginal tissue stores of vitamin A, might precipitate deficiency with serious consequences, including impairment of the immune response and of epithelial regeneration [30]. Our finding of a significant proportion of CF subjects with low serum vitamin A is in agreement with previous studies [3,4,27,31 – 33]. The normal increase in serum vitamin A with age was attenuated in CF subjects, who had significantly lower levels in adolescence and adulthood than controls. Vitamin A level was negatively associated with CRP, which increased with age in the CF group. This association of vitamin A and CRP is consistent with the work of Duggan et al. [34], who found that CRP was negatively correlated with vitamin A level in adult CF patients admitted for pulmonary exacerbation. Examination of the relationship between age and vitamin A separately in children, adolescents and adults provided an additional insight. There was an increasing disparity in mean vitamin A level between CF and control groups with age, suggesting an association with disease progression independent of pancreatic or intestinal malabsorption, in agreement with the findings of Lancelotti et al. [32], who found that serum vitamin A levels were not associated with pancreatic sufficiency status. The clinical significance of reduced serum vitamin A levels in CF is not known. We found that CF subjects with low serum vitamin A had lower levels of lung function, lower weight z-score and lower bone mineral density than those with normal vitamin A level, but any causal relationship, including interaction with inflammatory status, requires further investigation. The association with 1,25(OH)2D is a new observation. The highly significant negative relationship with 1,25(OH)2D, which did not differ in the regression with all subjects or that confined to CF subjects, may be related to immune modulatory and gene regulatory effects of both vitamin A and 1,25OH2D, which interact at the level of the retinoic acid receptor (RXR) and vitamin D receptor (VDR) [22]. Cawood et al. [35] using tracer techniques found that whilst post-prandial retinol was comparable in CF and control subjects, serum levels followed up from 7 to 28 days gradually declined in controls but were undetectable at any stage in the CF group. These results are indicative of reduced vitamin A mobilisation from the liver in spite of reduced serum levels. These data are supported by the known upregulation of hepatic stellate cells, responsible

148

R.M. Greer et al. / Journal of Cystic Fibrosis 3 (2004) 143–149

for secretion of the retinol binding protein (RBP) complex, in some patients with CF [36]. Our findings are also in agreement with the conclusions of Rosales et al. [10], who showed in rats that inflammation-induced hyporetinaemia was the result of reduced hepatic synthesis of retinol binding protein, and reduced secretion of the retinol binding protein – retinol complex from the liver. Rosales and Ross [37] further showed, in both vitamin A supplemented and nonsupplemented rats, that inflammation-induced hyporetinaemia represents tissue re-distribution of vitamin A rather than a loss of vitamin A. The presence of a hepatic transport defect, or defect of hepatic retinol release, is also supported by the work of Lepage et al. [38], who found improvement in the serum retinol/RBP ratio following ursodeoxycholic acid (URSO) treatment in children with CF. However Lindblad et al. [29] found that liver vitamin A level decreased with age in supplemented CF subjects aged 8 – 34 years, suggesting that low serum levels were not entirely attributable to a hepatic transport defect, and providing evidence for the safety of routine supplementation. Serum retinol binding protein, which may aid interpretation of serum vitamin A in the light of inflammatory status and liver function, was not measured and is a limitation of this study [24]. In addition, adherence with pancreatic enzyme replacement and/or vitamin supplementation was not assessed. Nevertheless serum vitamin A measurement is readily available, and currently the recommended tool to evaluate vitamin A status [39], but must be interpreted in the light of concurrent inflammatory status. The clinical questions raised by this study are whether low serum vitamin A levels in CF are due to nutritional vitamin A deficiency, the inflammatory response, or a combination of both. Wood et al. [40] found improvement in lung function in CF subjects receiving high-dose antioxidant supplementation. The role of the liver in vitamin A storage and transport in CF needs clarification. Studies are needed to determine the importance of serum vitamin A in the pathophysiology of lung disease in CF, and the potential role of supplementation (with appropriate surveillance for potential adverse effects), in the contexts of both chronic inflammatory disease and acute pulmonary exacerbation. Acknowledgements This study was supported by Cystic Fibrosis Australia (Australian Cystic Fibrosis Research Trust) and the Royal Children’s Hospital Foundation. The authors thank the study participants and their families and the staff of The Prince Charles Hospital and Royal Children’s Hospital. References [1] Greer RM, Buntain HM, Wainwright CE, et al. Vitamin A is associated with the inflammatory marker CRP in cystic fibrosis. J Cyst Fibros 2003;2(Suppl. 1):S87.

[2] Sinaasappel M, Stern M, Littlewood JM, et al. Nutrition in patients with cystic fibrosis: a European consensus. J Cyst Fibros 2002;1: 51 – 75. [3] Watkin S, Bell SC, Wynn S, et al. Vitamin A and E and pulmonary function in cystic fibrosis. Med Sci Res 1998;26:97 – 9. [4] Feranchak AP, Sontag MK, Wagener JS, et al. Prospective, long-term study of fat-soluble vitamin status in children with cystic fibrosis identified by newborn screen. J Pediatr 1999;135(5):601 – 10. [5] Sokol RJ. Vitamins A and E in cystic fibrosis. Pediatr Pulmonol 1993;9:167 – 8. [6] Louw JA, Werbeck A, Louw MEJ, et al. Blood vitamin concentrations during the acute-phase response. Crit Care Med 1992;20:934 – 41. [7] Thurnham DI, McCabe GP, Northrop-Clewes CA, et al. Effect of subclinical infection on plasma retinol concentrations and assessment of prevalence of vitamin A deficiency: meta-analysis. Lancet 2003; 362:2052 – 8. [8] Gillham B, Papachristodoulou D, Thomas J. Chapter 19. Nutrition: vitamins. In: Gillham B, Thomas JH, editors. Will’s Biochemical Basis of Medicine. Oxford: Butterworth Heinemann; 1997. p. 171 – 81. [9] Thurnham DI, Northrop-Clewes CA. Optimal nutrition: vitamin A and the carotenoids. Proc Nutr Soc 1999;58(2):449 – 57. [10] Rosales FJ, Ritter SJ, Zolfaghari R, et al. Effects of acute inflammation on plasma retinol, retinol-binding protein, and its messenger RNA in the liver and kidneys of vitamin A sufficient rats. J Lipid Res 1996;37:962 – 71. [11] Redlich CA, Rockwell S, Chung JS, et al. Vitamin A inhibits radiation-induced pneumonitis in rats. J Nutr 1998;128:1661 – 4. [12] Swamidas GP, Basaraba RJ, Baybutt RC. Dietary retinol inhibits inflammatory responses of rats treated with monocrotaline. J Nutr 1999;129:1285 – 90. [13] Reifen R. Vitamin A as an anti-inflammatory agent. Proc Nutr Soc 2002;61(3):397 – 400. [14] Neugebauer MA, Vernon SA, Brimlow G, et al. Nyctalopia and conjunctival xerosis indicating vitamin A deficiency in cystic fibrosis. Eye 1989;3(Pt. 3):360 – 4. [15] Huet F, Semama D, Maingueneau C, et al. Vitamin A deficiency and nocturnal vision in teenagers with cystic fibrosis. Eur J Pediatr 1997;156(12):949 – 51. [16] Glasziou P, Mackerras D. Vitamin A supplementation in infectious diseases: a meta-analysis. BMJ 1993;306(6874):366 – 70. [17] Berger M. Lung inflammation early in cystic fibrosis. Am J Respir Crit Care Med 2002;165:857 – 60. [18] Nixon GM, Armstrong DS, Carzino R, et al. Early airway infection, inflammation and lung function in cystic fibrosis. Arch Dis Child 2002;87(4):306 – 11. [19] Chmiel JF, Berger M, Konstan MW. The role of inflammation in the pathophysiology of CF lung disease. Clin Rev Allergy Immunol 2002;23(1):5 – 27. [20] Scott A, Matson A, Glanville A, et al. Changes in nutritional parameters following lung transplantation for end stage cystic fibrosis. Respirology 2002;7:A46 (Supplement). [21] Greer RM, Buntain HM, Potter JM, et al. Abnormalities of the PTH – vitamin D axis and bone turnover markers in children, adolescents and adults with cystic fibrosis: comparison with healthy controls. Osteoporos Int 1989;14(5):404 – 11. [22] Jimenez-Lara A, Aranda A. Interaction of vitamin D and retinoid receptors on regulation of gene expression. Horm Res 2000;54(5 – 6):301 – 5. [23] Stephensen CB, Gildengorin G. Serum retinol, the acute phase response, and the apparent misclassification of vitamin A status in the third National Health and Nutrition Examination Survey. Am J Clin Nutr 2000;72(5):1170 – 8. [24] Schweigert F. Inflammation-induced changes in the nutritional biomarkers serum retinol and carotenoids. Curr Opin Clin Nutr Metab Care 2001;4:477 – 81. [25] Shosayov D, Bibi H, Biesalksi H, et al. Repeated allergen challenge in rats increases vitamin A consumption. Chest 2002; 122(4):1407 – 11.

R.M. Greer et al. / Journal of Cystic Fibrosis 3 (2004) 143–149 [26] Mitra AK, Alvarez JO, Stephensen CB. Increased urinary retinol loss in children with severe infections. Lancet 1998;351(9108):1033 – 4. [27] Sokol RJ, Reardon MC, Accurso FJ, et al. Fat-soluble vitamins in infants identified by cystic fibrosis newborn screening. Pediatr Pulmonol Suppl 1991;752 – 5. [28] Abbott JDMBD, et al. Treatment compliance in adults with cystic fibrosis. Thorax 1994;49:115 – 20. [29] Lindblad A, Diczfalusy U, Hultcrantz R, et al. Vitamin A concentration in the liver decreases with age in patients with cystic fibrosis. J Pediatr Gastroenterol Nutr 1997;24(3):264 – 70. [30] Stephensen CB. Vitamin A, infection, and immune function. Annu Rev Nutr 2001;21:167 – 92. [31] Kawchak DA, Sowell AL, Hofley PM, et al. Longitudinal analysis shows serum carotenoid concentrations are low in children with cystic fibrosis. J Am Diet Assoc 1999;99(12):1569 – 72. [32] Lancellotti L, D’Orazio C, Mastella G, et al. 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(4):281 – 5. [33] Madarasi A, Lugassi A, Greiner E, et al. Antioxidant status in patients with cystic fibrosis. Ann Nutr Metab 2000;44(5 – 6):207 – 11.

149

[34] Duggan C, Colin AA, Agil A, et al. Vitamin A status in acute exacerbations of cystic fibrosis. Am J Clin Nutr 1996;64(4):635 – 9. [35] Cawood A, Murphy J, Carroll M, et al. Altered availability and mobilisation of vitamin A in CF. 26th European Cystic Fibrosis Conference Belfast: European Cystic Fibrosis Society. J Cyst Fibr 2003; A341(Suppl. 2):85. [36] Lewindon PJ, Pererra TN, Hoskins AC, et al. The role of hepatic stellate cells and transforming growth factor-beta(1) in cystic fibrosis liver disease. Am J Pathol 2002;160(5):1705 – 15. [37] Rosales FJ, Ross AC. Acute inflammation induces hyporetinemia and modifies the plasma and tissue response to vitamin A supplementation in marginally vitamin A-deficient rats. J Nutr 1998;128:960 – 6. [38] Lepage G, Paradis K, Lacaille F, et al. Ursodeoxycholic acid improves the hepatic metabolism of essential fatty acids and retinol in children with cystic fibrosis. J Pediatr 1997;130(1):52 – 8. [39] Bell SC, Shepherd R. Optimising nutrition in cystic fibrosis. J Cyst Fibros 2002;1:47 – 50. [40] Wood LG, Fitzgerald DA, Lee AK, et al. Improved antioxidant and fatty acid status of patients with cystic fibrosis after antioxidant supplementation is linked to improved lung function. Am J Clin Nutr 2003;77(1):150 – 9.