Retinol protects respiratory epithelial cells against oxidation of endogenous or exogenous origin. This is a particularly valuable role for patients with cystic fibrosis (CF) who must make a continued effort against their primary oxidative stress, worsened by chronic respiratory inflammation, and whose prognosis is dependent on pulmonary function (1,2). Unfortunately, because of limited inputs and high consumption and losses, vitamin A deficiency is frequent among these patients from an early age (3). Although supplementation can achieve a normal concentration of serum retinol (SR) (1), its antioxidant power remains insufficient, or even depleted, thus allowing epithelial respiratory damage (4).
Warning of the association of SR with clinical evolution and lung function in patients with CF, a European Consensus published in 2002 (5) addressed the question as to whether the final goal of antioxidant supplementation is to achieve higher plasma levels in subjects with CF than in healthy people. It was subsequently shown that patients with low levels of SR, although within the normal range, experience more frequent pulmonary exacerbations (6). This seems to support a recently suggested level of relation between vitamin A status and clinical evolution in patients with CF (7), far exceeding the simple nutritional goal of restoring normal serum values. This crucial issue demands elucidation and a well-defined response.
The study goal was to explore the relation of SR with lung function by means of a suitable sample of young patients with CF, prospectively excluding even subclinical inflammatory conditions and all of the patients with vitamin A deficiency.
This is a prospective cross-sectional study on children and youth with CF, conducted in the CF clinic from 5 tertiary care Spanish hospitals. For 1 year, we selected young patients with CF quarterly monitored and clinically stable (no cough or dyspnea), with the ability to perform reproducible spirometry.
In a first screening, all of those deficient in vitamin A (SR <20 g/dL) were excluded, as were those at risk for retinol toxicity, such as protein malnutrition, liver disease, hyperlipidemia, alcohol intake, pregnancy, or voriconazol treatment (8,9). This way, an initial group of 124 patients was preselected, all of whom provided written informed consent. The study was approved by the institutional ethics committee at each participating center.
The second visit, one quarter thereafter, included clinical history (inquiring as to anorexia, weakness, headache, arthralgia, fever, and cough) and examination (xerosis, xerophthalmia, dyspnea, or new crackles) to identify pulmonary exacerbation (10) and vitamin A deficiency or toxicity (11). Thereby the group was reduced to 104 cases. SR (high-performance liquid chromatography (12)), serum calcium, cholesterol, triacylglycerol, aspartate aminotransferase, alanine aminotransferase, γ-glutamyl transferase, alkaline phosphatase, C-reactive protein (CRP), and fecal elastase-1 (monoclonal enzyme-linked immunosorbent assay test) were determined in all of these cases.
The final study group included 98 patients with CF with SR >20 μg/dL, which were free of pulmonary exacerbation (CRP up to 0.5 mg/dL) and hepatic dysfunction (serum enzymes lower than twice the upper limit of normal). All of them meet the CF diagnostic criteria (13) with a verified sweat chloride test. Cystic fibrosis transmembrane conductance regulator mutation was identified in all but 2 patients (whose sweat chloride tests were 97 and 102 mmol/L, respectively).
Height, weight, and the computed body mass index (BMI) were standardized (Zheight, Zweight, ZBMI) for Spanish population references (14). As normal SR values are age-stratified and our cohort extends in a wide range of ages, SR values were adjusted to the distribution in healthy population (National Health and Nutrition Examination Survey [NHANES] study) (15). Spirometry was performed as required by the American Thoracic Society/European Respiratory Society Task Force (16), expressing lung function indices (forced expiratory volume in 1 second [FEV1], forced vital capacity [FVC], and forced expiratory flow between 25% and 75% of FVC [FEF25–75]) as a percentage of predicted normal.
Values are expressed as mean ± standard deviation score (SDS) or proportions. Normality of the distribution was assessed using Kolmogorov-Smirnov and Shapiro-Wilk tests. Student t test and one-way analysis of variance with Bonferroni post hoc tests were used for the comparison of means. Mann-Whitney U test was used when t tests were not applicable. Pearson χ2 test was used for categorical variables and odds ratio (OR) was obtained by logistic regression analysis. Pearson correlation coefficient was applied when assessing the linear association between numerical variables. To avoid confusion, adjusted coefficients for sex, age, ZBMI, and pancreatic function were obtained via multiple linear regression analysis. The Jonckheere-Terpstra trend test, a nonparametric test for ordered differences among ordinally arranged groups, was performed to assess linear trend relations. P values <0.05 were considered statistically significant. Analysis was performed using the Statistical Package for the Social Sciences (SPSS Inc, Chicago, IL) version 15.0 for Windows.
The cohort of 98 patients, from 6.8 to 22.3 years of age, was composed of 32 children up to 11 years, 56 adolescents between 12 and 18 years, and 10 adults. Table 1 includes most of their descriptive data. Eighty-seven with <200 mg/g stool of fecal elastase-1 (17) constituted the pancreatic insufficient (PI) set. They were receiving 3954 ± 1968 IU/kg of lipase, and 978 ± 346 μg of preformed retinol daily, a dosage adjustment looking for a SR normal range (5). Despite this, patients up to 11 years of age received a similar amount of retinol but higher relative dosage than those older (32.1 ± 12.0 vs 20.2 ± 9.5 μg/kg weight/day) and achieved higher ZSR (SR z score) (2.31 ± 2.02 vs 0.99 ± 1.83 SDS).
The remaining 11 patients, whose fecal elastase-1 was normal, constitute the pancreatic sufficient (PS) set. All but 3 of them received a discretionary supplement of vitamin A (658 ± 111 μg/day).
Zheight (−0.12 ± 1.29 SDS in the total group) is markedly higher in males (0.21 ± 1.38 vs −0.40 ± 1.15 SDS; P < 0.05) and could be higher in PS (0.52 ± 1.50 SDS) than in PI patients (−0.25 ± 1.23 SDS; P = 0.07). There were no differences between the Zheight of age groups determined by puberty.
Zweight of the whole group was −0.72 ± 0.78 SDS, and ZBMI was −0.38 ± 0.83 SDS. Only 29 patients were adequately nourished (BMI in the 50th percentile or higher (18)), whereas BMI was lower than the 10th percentile (ie, nutritional failure) in 16 patients. Both Zweight and ZBMI were higher in the PS set (−0.03 ± 1.02 and 0.37 ± 0.90 SDS) than in the PI set of patients (−0.80 ± 0.71 and −0.48 ± 0.77 SDS, respectively) (P < 0.005). No sex differences were related to the nutritional status.
Serum cholesterol, triacylglycerol, aspartate aminotransferase, alanine aminotransferase, γ-glutamyl transferase, and alkaline phosphatase fell within the normal range (as did CRP by selective criterion), with neither sexual dimorphism nor age-related differences in any of them.
SR of the study group was 56.6 ± 18.4 μg/dL, and ZSR 1.38 ± 1.96 SDS. The whole group was above the 2.5th percentile (Table 1), with 31 cases >97.5th percentile (54.4 μg/dL in patients under 12 years, 7.2 μg/dL for those between 12 to 19 years, and 89.1 μg/dL for all of 20 or more) (highest value 110 μg/dL). Mean age of this subgroup of patients with SR above the normal reference range (SRA) was lower than that of those with SR within the normal range (SRN) (Table 1), with a proportionally higher share of children up to 11 years (14, 56% of them) than older patients (17, 23% of them) (χ2 < 0.005); however, the proportion of PI patients (25; 29% of them) and PS patients (6; 54% of them) was not significantly different (χ2 = 0.3), nor were different ZBMI or the proportion of patients in nutritional failure or within the nutritional goal for patients with CF (Table 1). Neither sex proportions were different in SRA and SRN subgroups.
SR in the PS set was higher than that in the PI set (67.4 ± 19.0 vs 55.5 ± 18.3 μg/dL; P < 0.05), finding similar difference in ZSR (2.54 ± 2.50 vs 1.12 ± 1.81 SDS; P < 0.05). No correlation was found between retinol supplementation and SR or any of the respiratory function indexes, nor in the total group nor in the PI set (r = 0.04), and supplementation of PI patients of the SRA subgroup was similar to that in the SRN subgroup (25.6 ± 11.8 vs 24.1 ± 10.6 μg/kg body weight; P = 0.56); however, in PS patients, retinol supplementation is highly correlated with SR level (r = 0.777; P < 0.01), FEV1 (r = 0.850; P < 0.05), and FVC (r = 0.855; P = 0.01), and retinol supplementation in PS patients of the SRA subgroup was greater than in those of the SRN subgroup (17.5 ± 4.9 vs 6.8 ± 7.9 μg/kg body weight; P < 0.05).
Respiratory Function Indices
Respiratory functional data in the overall group were widely scattered: 36% to 126% of predicted values for FEV1, 46% to 130 % for FVC, and 21% to 124% for FEF25–75. None of them showed significant differences between patients up to 11 years and the older ones. Boys had slightly better data than girls, but only in FVC there were significant differences (96.4% ± 11.9% vs 89.4% ± 17.9%; P < 0.05).
Patients in nutritional failure had the same FEV1 (84.3 ± 18.0 vs 89.4% ± 16.2%; P > 0.05) and FEF25–75 (55.7% ± 20.0% vs 64.0% ± 20.1%; P > 0.05), although poorer FVC (88.1% ± 17.0% vs 95.0% ± 14.6%; P < 0.05) compared with the rest. None of the respiratory indices in patients with BMI >50th percentile was significantly better than that in the rest. FEV1 and FVC of PS and PI sets were similar (Table 1), although FEF25–75 was higher in the former (71.6 ± 28.8 vs 56.2 ± 22.9; P < 0.05).
Serum Retinol and Respiratory Function Indices
Raw SR correlates positively with FEV1 (r = 0.318; P = 0.001), FVC (r = 0.237; P < 0.05), and FEF25–75 (r = 0.287; P < 0.05). So does ZSR, even more strongly, with FEV1 (r = 0.364; P < 0.001), FVC (r = 0.203; P < 0.05), and FEF25–75 (r = 0.296; P < 0.01). These associations are adjusted for sex, age, BMI, and pancreatic function (B = 0.210; P = 0.028). Figure 1 shows this correlation in the total group and in the PS and PI sets of patients.
In the PI set, raw SR also correlates with FEV1 (r = 0.282; P = 0.01), as does ZSR (r = 0.353; P = 0.001). Raw SR of PS patients strongly correlates with FEV1 (r = 0.712; P < 0.05), although correlation of ZSR with FEV1, in this small set, is on the edge of significance (r = 0.589; P = 0.059).
FEV1 of SRN subgroup patients in the lower half of the NHANES distribution (from 2.5th to 50th percentile) did not differ from that of those in the upper half (from 50th to 97.5th percentile) (80.2% ± 17.5% vs 87.4% ± 17.3%; P > 0.05); however, FEV1 of the SRA subgroup significantly outperforms that of the SRN subgroup (Table 1), in the whole group and in both PI and PS sets (Fig. 2). Among these patients this difference is also established by FVC and FEF25–75.
FEV1 is <80% predicted in 34.2% (23/67) of cases in SRN subgroup, versus the 9.7% (3/31) of those of the SRA subgroup (χ2 < 0.001). No case in the latter had FEV1 <70% predicted; however, FEV1 was >100% in 16.4% (11/67) of the SRN subgroup, versus 32% (10/31) of the SRA (χ2 < 0.001).
The OR of having FEV1 >80% in SRN cases located in the upper half of the normal range of SR is similar to that in the lower half (OR 1.04 vs 0.96); however, the OR of having a FEV1 >80% in SRA patients is 3.78 (probability 79%) versus 0.26 (probability 21%) in SRN patients.
The Jonckheere-Terpstra test applied to these 3 SR classes of patients (lower half and upper half of SRN subgroup and SRA subgroup) gives a significance coefficient of 0.003. The coefficient for the PI and PS sets was 0.02 and 0.03, respectively. This means that both in the overall group and in each of the PI and PS sets, there is a significant upward linear trend of FEV1, related to SR, with a noteworthy role of the 31 cases of the SRA subgroup (Fig. 2).
Tolerance to Retinol
No signs or symptoms of retinol toxicity were found in the clinical monitoring of our patients. The Student t test found no differences in serum cholesterol, triacylglycerol, transferases, or alkaline phosphatase between SRA and SRN subgroups of patients. Serum calcium was slightly higher in SRA than in SRN subgroup (9.73 ± 0.43 vs 9.45 ± 0.43 mg/dL, P < 0.01), although with no clinical relevance.
Pulmonary integrity is crucial for the evolution of patients with CF; hence, their prognosis depends on its preservation. Vitamin A plays an active role in epithelial protection, preserving against oxidative damage and participating in the immune response. Its deficiency, related to insufficient intake or inflammation, translates into pulmonary deterioration (11). Therefore, knowledge and control of its effects are particularly important for the management and survival of these patients.
After the report by Carr and Dinwiddie (19) on the strong association between low SR (<9 μg/mL) and poor lung function in their mostly vitamin A–deficient group of patients, it was retrospectively found that higher levels of SR within the normal range correlate with higher recorded values of FEV1(20); however, although SR lower than 20 μg/dL evidences vitamin A deficiency and low liver stores, a normal SR in itself does not ensure its sufficiency (11). Actually, their oxidative imbalance persists despite normal SR (4,21), whereas only supranormal SR level can ensure full organic stores (11). It could thus be inferred that a moderately high level of SR ensures the greatest antioxidant power related to this factor.
With the aim of investigating this issue, all of the patients with vitamin A deficiency (SR <20 μg/dL) were excluded from the present study. Moreover, as SR decreases sharply during inflammation, even subclinical (22,23), a fast acute-phase reactant was used to achieve early detection of pulmonary exacerbations. All of the known conditions of singular retinol toxicity risk (8,9) were preemptively excluded as well. Besides, SR values were adjusted to the mean for age of healthy people, which is clearly age-stratified (15). Furthermore, lipase supplementation was likewise controlled because of the importance of steatorrhea on the intestinal absorption of this liposoluble vitamin.
After controlling for confounding factors, our study group showed a positive relation between SR and respiratory functional indices. The SR typified for the patient's age determines 13.6% of the variability of FEV1 in the entire study group (11% for the PI set; 34.6% for the PS). This positive correlation is even more marked in patients with moderately high SR levels (group SRA), which preserve the best respiratory function indices of the cohort studied (>90% had FEV1 ≥80%; none with FEV1 <70%).
Patients with CF experience a harmful oxidative status, despite being within the normal SR range (4), because their needs of antioxidants, such as retinol, are higher than those of the healthy population. In our patients, this means that in 1 of 3 of those with normal SR (SRN subgroup) FEV1 is <80% predicted. Therefore, at that level, they do not receive sufficient support for their health care needs.
Consistently, retinol prescription to patients with CF must be regarded as an antioxidant agent, in addition to as a nutritional supplement. As seen in our experience, SR slightly above the upper NHANES limit noticeably improves their respiratory function and implicitly their life expectancy (Fig. 3). It is necessary to know whether this target can be achieved with enough security margin in regard to toxicity related to retinol excess.
The present goal for vitamin A supplementation in patients with CF should be high enough so as to achieve a normal serum level, without provoking adverse effects (5), with a consensuated dosage (up to 3400 μg of retinol activity equivalents since 8 years of age) (5,24), several times higher than the recommended dietary allowance (from 600 for 9- to 13-year-old schoolchildren to 900 μg/day for adults) (9). With that goal, and taking into account the risk of toxicity related to retinol excess (5), our patients were receiving a modest dose (lower than that of other study groups (6,25)), although about 30% of them had an SR mildly above the upper NHANES limit (15); however, all were lower than the “no adverse effect level observed,” far below the reported toxic range (9). None of our cases showed undesirable clinical effects in any of the investigated aspects, nor were abnormal data reported in any of the ad hoc analysis included. Actually, SR levels up to 110 μg/dL are within the normal range for other broader references (6,25) (albeit we feel they suit poorly to the ages of the patients included).
It was proposed to increase the target SR to at least 35 μg/dL (6), which is the mean level for healthy children up to 11 years (15). Indeed, this is a modest proposition that may be scanty for them, and even more for older patients (with higher reference values) because it does not warrant their antioxidant needs.
Inefficient and disparate retinol absorption and metabolism in PI individuals (18) hinder a proportionate serum level response to vitamin A supplementation; however, almost 1 in 3 of our patients, who were receiving a moderate retinol supplementation, exceeded the upper limit of normal. Still it is expected that, as already reported (26), a higher dose (although within the recommended range) may lead to supranormal serum levels, especially during inflammation-free periods. Indeed, SR downfall during inflammation is related to consumption and to metabolic blocking (23), so that when the acute episode is solved, the SR can rise faster than in cases of deficiency. As stated, it is necessary to recognize those cases of low SR related to an inflammatory response, a dissimilar condition to a nutritional deficiency with empty stores (23). This means that dosage should be individualized for each patient (age, body size, pancreatic function) and their present needs: a moderate, individually adjusted dose along infection-free periods; and higher, within the wide margin recommended, to meet the increased requirements during exacerbations.
In view of both the actual antioxidant necessities, as the latent toxicity risk of the vitamin-supplemented patients with CF, such a therapeutic approach requires the always advisable periodic serum monitoring (5), even quarterly if necessary. Fortunately, a vigorous supplementation with β-carotene achieves the needed antioxidant power, avoiding those retinol potential risks (27).
Evaluation of the relation between SR and lung function must always take into account that although retinol levels can change substantially in a few weeks, a pulmonary functional loss related to chronic bronchopulmonary inflammation has no such reversibility. A longitudinal study with a controlled cohort of young patients without significant lung injury would probably provide a true picture of the extent of the benefits of retinol.
Lung function of young patients with CF correlates directly with SR, this being especially noteworthy in those with a moderately high retinol level (up to 110 μg/dL). This subgroup maintains the best respiratory function (FEV1 >80% in >90% of them). This moderate rise in retinol does not cause any signs of toxicity and is far from its toxic level. The aim of maintaining SR level within the normal limits must be revised to achieve better health care.
1. Brown RK, Wyatt H, Price JF, et al. Pulmonary dysfunction in cystic fibrosis is associated with oxidative stress. Eur Respir J
2. Madarasi A, Lugassi A, Greiner E, et al. Antioxidant status in patients with cystic fibrosis. Ann Nutr Metab
3. Ahmed F, Ellis J, Murphy J, et al. Excessive faecal losses of vitamin A (retinol) in cystic fibrosis. Arch Dis Child
4. Lagrange-Puget M, Durieu I, Ecochard R, et al. Longitudinal study of oxidative status in 312 cystic fibrosis patients in stable state and during bronchial exacerbation. Pediatr Pulmonol
5. Sinaasappel M, Stern M, Littlewood J, et al. Nutrition in patients with cystic fibrosis: a European Consensus. J Cyst Fibros
6. Hakim F, Kerem E, Rivlin J, et al. Vitamins A and E and pulmonary exacerbations in patients with cystic fibrosis. J Pediatr Gastroenterol Nutr
7. Munck A. Nutritional considerations in patients with cystic fibrosis. Expert Rev Resp Med
8. Cheng MP, Paquette K, Lands LC, et al. Voriconazole inhibition of vitamin A metabolism: are adverse events increased in cystic fibrosis patients? Pediatr Pulmonol
9. Institute of Medicine. Food and Nutrition Board. Dietary Reference Intakes for Vitamin A, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc
. Washington, DC: National Academies Press; 2001:82–161.
10. Bilton D, Canny G, Conway S, et al. Pulmonary exacerbation: towards a definition for use in clinical trials. Report from the EuroCareCF Working Group on outcome parameters in clinical trials. J Cyst Fibros
2011; 10 (suppl 2):S79–S81.
11. Nakajoh M, Fukushima T, Suzuki T, et al. Retinoic acid inhibits elastase-induced injury in human lung epithelial cell lines. Am J Respir Cell Mol Biol
12. Quesada JM, Mata-Granados JM, Luque de Castro MD. Automated method for the determination of fat-soluble vitamins in serum. J Steroid Biochem Mol Biol
13. Farrell PM, Rosenstein BJ, White TB, et al. Guidelines for diagnosis of cystic fibrosis in newborns through older adults: Cystic Fibrosis Foundation consensus report. J Pediatr
14. Sobradillo B, Aguirre A, Aresti U, et al. Curvas y tablas de crecimiento (Estudios longitudinal y transversal). In: Fundación Faustino Orbegozo Eizaguirre
. Majadahonda (Madrid): Ergon; 2004.
16. Miller MR, Hankinson J, Brusasco V, et al. Standardisation of spirometry. Eur Respir J
17. Leus J, Van Biervliet S, Robberecht E. Detection and follow up of exocrine pancreatic insufficiency in cystic fibrosis: a review. Eur J Pediatr
18. Stallings VA, Stark LJ, Robinson KA, et al. Clinical Practice Guidelines on Growth and Nutrition Subcommittee; Ad Hoc Working Group. Evidence-based practice recommendations for nutrition-related management of children and adults with cystic fibrosis and pancreatic insufficiency: results of a systematic review. J Am Diet Assoc
19. Carr SB, Dinwiddie R. Vitamin A as a predictor for lung function in cystic fibrosis [Abstract]. Pediatr Pulmonol
1996; 13 (suppl):317.
20. McKeever TM, Lewis SA, Smit HA, et al. A multivariate analysis of serum nutrient levels and lung function. Respir Res
21. Wood LG, Fitzgerald DA, Gibson PG, et al. Oxidative stress in cystic fibrosis: dietary and metabolic factors. J Am Coll Nutr
22. Thurnham DI, McCabe GP, Northrop-Clewes CA, et al. Effects of subclinical infection on plasma retinol concentrations and assessment of prevalence of vitamin A deficiency: meta-analysis. Lancet
23. Greer RM, Buntain HM, Lewindon PJ, et al. Vitamin A levels in patients with CF are influenced by the inflammatory response. J Cyst Fibros
24. Borowitz D, Baker RD, Stallings V. Consensus report on nutrition for pediatric patients with cystic fibrosis. J Pediatr Gastroenterol Nutr
25. Aird FK, Greene SA, Ogston SA, et al. Vitamin A and lung function in CF. J Cyst Fibros
26. Graham-Maar RC, Schall JI, Stettler N, et al. Elevated vitamin A intake and serum retinol in preadolescent children with cystic fibrosis. Am J Clin Nutr
27. Sagel SD, Sontag MK, Anthony MM, et al. Effect of an antioxidant-rich multivitamin supplement in cystic fibrosis. J Cyst Fibros