Main

A high prevalence of skeletal lesions, such as osteoporosis and kyphosis, and increased fracture rate have been reported in patients with CF(1–9). Pathogenetic factors include malnutrition, vitamin D and calcium malabsorption, pulmonary acidosis, poor sun exposure, and relative inactivity(2–5, 7, 9). In older patients, delayed puberty or hypogonadism may play a pathogenetic role, too(4, 10, 11). Due to the increasing life expectancy of CF patients, more attention needs to be given to the effects of the disease on bone turnover and bone mineral status.

Recently, some biochemical markers have been proposed to provide information about the dynamics of bone turnover. OC or boneγ-carboxyglutamic acid protein, a major noncollagenous protein of the bone matrix specifically secreted by osteoblasts, is considered a valid marker for the examination of bone formation, in that its serum levels are higher at ages when the bone mineralization rate is increased(12).

Type I collagen, the most abundant body collagen, is a major product of osteoblasts, accounting for more than 90% of the organic bone matrix(13). Type I collagen is also the major extracellular matrix protein of soft tissues, where it always occurs together with other collagens, mainly with the fibrillar type III collagen(13, 14). Types I and III collagens are derived from specific precursor molecules called procollagens, which are synthesized intracellularly. During the synthesis, large and soluble propeptide domains are released into the circulation from the precursor molecules in a stoichiometric ratio of 1:1(14). Serum PICP concentrations are generally related to the rate of bone formation(15); whereas serum PIIINP concentrations seem to be a sensitive early marker of developing fibrotic process in conditions in which there is a pathologic accumulation of type III collagen, such as chronic liver(16, 17) or lung(18, 19) diseases, and in patients with diabetic microangiopathy(20, 21). Serum ICTP and urinary cross-linked NTX, which are breakdown products of mature cross-linking in bone collagen, seem to reflect the bone resorption rate(22–24).

During childhood, serum levels of OC(25), PICP(26), PIIINP(27), ICTP(28), and NTX(24) are much higher than in adulthood, with patterns resembling growth velocity curves. Therefore, these biochemical markers also reflect somatic growth rate, in addition to bone turnover and collagen metabolism.

The purpose of this study was to measure OC, PICP, PIIINP, ICTP, and NTX, and TBBM in prepubertal, pubertal, and young adult CF patients to 1) investigate the dynamics of bone turnover and collagen metabolism,2) evaluate bone mineral status at different ages, and 3) assess whether the measurement of these biochemical markers may represent a useful tool in the management of these patients.

METHODS

Patients. A total of 59 CF patients (32 male and 27 female) receiving regular outpatient clinical care at CF Center of Messina University Hospital were examined. Patients were subdivided as prepubertal, pubertal, and young adults according to their chronologic age and pubertal development(Table 1). Shwachman-Kulczycki score(29) and Chrispin-Norman score(30) indicated a higher degree of disease severity in young adult patients than in prepubertal and pubertal patients(Table 1). Pulmonary function tests, including forced vital capacity, forced expiratory volume in 1 s, and forced expiratory volume in 1 s/forced vital capacity ratio, as well as arterial pH, did not differ among the groups, even though they were slightly but not significantly lower in young adult patients than in prepubertal and pubertal patients(Table 1). Diagnosis of CF was made on clinical findings, positive family history, and a sweat chloride concentration higher than 60 mmol/L, determined by pilocarpine iontophoresis. Exocrine pancreatic insufficiency was present in all patients, as assessed by pathologic fecal fat losses. In all patients, serum alanine and aspartate aminotransferases,γ-glutamyltranspeptidase, total alkaline phosphatase, total and direct bilirubin, and creatinine levels were below the upper limit of our laboratory reference range (data not shown). Moreover, all patients had normal serum concentrations of calcium, phosphate, and 25-hydroxyvitamin D (data not shown). Two young adult patients (1 male and 1 female) had insulin-dependent diabetes mellitus; they also showed ophthalmologic findings of diabetic microangiopathy (increased venous distension and tortuosity, retinal hemorrhage, and microaneurisms), whereas no sign of nephropathy or neuropathy was evident. Liver ultrasonography showed increased size (n = 8), increased size associated with dishomogenous echogenicity (n = 4), hyperechogenic parenchyma (n = 2), portal hypertension (n= 4), and gallstones (n = 2) or sludge (n = 1). The remaining patients (n = 38) had ultrasonographic normal livers. No patient underwent liver biopsy. No biochemical sign of hypogonadism was present in the patients, i.e. normal values of LH and FSH (basal and gonadotropin-releasing hormone-stimulated, 100 μg i.v.) and gonadal steroids for their pubertal stage.

Table 1 Clinical findings of subjects*
Full size table

Dietary management consisted in high caloric food intake (approximately 150% of recommended dietary allowances)(31). All patients received high dose supplementation of an enteric coated enzyme preparation (Pancrease, Cilag), vitamin A and vitamin D2 supplement at the dose of 1500 μg retinol equivalents/daily, and 20 μg/daily (800 IU/daily), respectively (Protovit Rafforzato, Roche), and vitamin E at a dose of 200 mg of α-tocopherol equivalents/daily (Evion Forte, Bracco). Furthermore, a calcium supplement of 728 ± 390 mg/daily, 906 ± 412 mg/daily, and 1150 ± 539 mg/daily was administered in prepubertal, pubertal, and young adult patients, respectively. Twenty patients(prepubertal, n = 8; pubertal, n = 3; young adults,n = 9) also received inhaled beclomethasone dipropionate (Clenil A, Chiesi) at a dose of 980 ± 200 μg/daily; none used oral glucocorticoids during the 3 mo preceding the study, or was taking thiazide diuretics. No patient performed physical exertion beyond that of daily normal activities. Compliance with therapy was good in all patients.

Controls. The control group consisted of 72 normal-statured, healthy subjects (38 male and 34 female) of similar age, sex, and BMI subdivided as prepubertal, pubertal, and young adults (Table 1). The range of maturity of the pubertal group of controls was similar to that of the pubertal group of patients. Pubertal stage was appropriate for chronologic age in all the control subjects, whereas it was delayed in some patients (4 pubertal and 2 young adult male subjects, and 2 pubertal and 1 young adult female subject) (Table 1).

Study design. In all patients and control subjects, serum concentrations of OC, PICP, PIIINP, and ICTP, and urinary values of NTX were measured after an overnight fast (between 0800 and 0900 h). At the time of evaluation of the biochemical markers, TBBM was also assessed in all but four prepubertal patients. Prevalence of bone fracture was investigated in both patients and control subjects; information obtained included the location of the fracture, the mechanism of injury, and the functional outcome.

Informed consent to perform the study was obtained from the parents of each subject when the chronologic age of their child was lower than 18, and directly from each subject when the chronologic age was more than 18. The study was approved by the ethics committee for human investigation of both University Hospital of Messina and Department of Pediatrics of University of Pisa.

Assessment of anthropometric findings. Standing height was measured with a wall-mounted stadiometer. To allow a comparison between different ages and genders, height was expressed as z score with respect to height SD according to the method of Tanner et al.(32) by using the formula: measured individual value - mean normal value for age and gender/SD of normal mean. Pubertal stage was assessed according to Tanner and Whitehouse(33). Bone age was evaluated by using the Greulich and Pyle method(34). BMI was calculated using the formula weight(kg)/height (m2).

Assessment of disease severity and liver ultrasonography. Patient's clinical and roentgenographic status was assessed by Shwachman-Kulczycki score (total scores ranged from 20 to 100, with a higher score indicating better functioning and less disease severity)(29), Chrispin-Norman score (total scores ranged from 0 to 38, with a lower score indicating a lesser degree of pulmonary disease)(30), and pulmonary function tests. Forced vital capacity and forced expiratory volume in 1 s were evaluated by a spirometer(Vitalograph, Inc., Lenexa, KS). Criteria for liver ultrasound interpretation were as follows. Fatty liver was suspected in the presence of hyperechogenic parenchyma, cirrhosis in the presence of heterogeneity and nodules, and portal hypertension in the presence of splenomegaly and dilated collateral veins, as proposed by Colombo et al.(35). Standard criteria for the gallbladder were used: decreased size for microbladder, and increased echoes and acoustic shadows for lithiasis or sludge.

Assessment of biochemical markers. Serum samples were separated within 2 h from sampling and stored at -20°C until assayed. Urine samples(second voluntary voiding) were stored at -20°C until analyzed. Serum concentrations of OC were measured by a two-site immunoradiometric assay, using a commercial kit (Human OC, Nichols Inst., San Juan Capistrano, CA), that recognized both intact OC and its large N-terminal mid-fragment. Serum PICP, PIIINP, and ICTP concentrations were detected by RIA using commercial kits (Orion Diagn., Espoo, Finland). The PICP antigen in serum had the same molecular size and the same affinity for the antibodies as standard PICP(28). The method for the detection of PIIINP selectively measured the intact PIIINP and high molecular weight antigen forms of PIIINP(36). The method to assess ICTP used polyclonal antibodies developed in rabbits and was based on a complex peptide that contained material from three polypeptide chains, two of them being derived from the carboxy-terminal telopeptide of one type I collagen molecule and the third from the triple-helical region of another molecule, plus a trivalent cross-link(22, 28). Urinary NTX values were measured by ELISA (Osteomark, Ostex Int., Seattle, WA) using a mouse MAb that specifically recognized urinary trivalent cross-linked peptides formed between collagen type I amino-terminal telopeptides and a lisine or hydroxylisine residue from a collagen triple helical domain(37). NTX values were expressed as equivalents per mole of bone type I collagen (BCE); the BCE values were expressed normalized to urinary creatinine (nmol/mmol) measured in the same urine sample. Both ICTP and NTX derive only from collagen degradation and are specific for type I collagen(28). For OC, interassay and intraassay coefficients of variation and sensitivity were 6.3%, 5.2%, and 0.05 μg/L, respectively; for PICP, 4.7%, 3.5%, and 1.2μg/L; for PIIINP, 6.5%, 5.2%, and 0.2 μg/L; for ICTP, 6.8%, 5.0%, and 0.5 μg/L; for NTX, 7.4%, 6.0%, and 20 nmol BCE. All blood and urine samples were measured in duplicate.

Assessment of TBBM. TBBM values were determined by dual energy x-ray absorptiometry using a Norland XR-26 (Norland Co., Fort Atkinson, WI). The total body scan was performed with the subject positioned supine on the scanning table; the same details for subject positioning suggested by Faulkner et al.(38) were used. The major component of TBBM is cortical bone (80% versus 20% of trabecular bone)(39). TBBM values were compared with appropriate sex-and age-reference values by using normative data of Rico et al.(39) in prepubertal and pubertal patients and those of Rico et al.(40) in young adult patients. The results were calculated as the z score by using the same formula we used to calculate height z score. The radiation dose to the patient was less than 5 mrem. The in vivo coefficient of variation was 1.4%

Statistical analysis. The results are expressed as X± SD. Comparison of the data were determined with the nonparametric Wilcoxon's (Mann-Whitney) rank-sum test by using a statistical system(LabStat. 303™, SIBIOC, Milan) adapted for IBM personal computer. Linear regression analysis by Pearson's formula was performed to determine r values. A p < 0.05 was considered significant.

RESULTS

Height, biochemical markers, and TBBM in prepubertal patients. Mean height was moderately reduced (-0.6 ± 0.7 z score); the deficit in height was more evident in male subjects than in female subjects(-0.9 ± 0.7 z score and -0.2 ± 0.6 z score, respectively, p < 0.02). Mean serum OC, PICP, and PIIINP concentrations did not differ from controls(Figs. 1,2,and 3, respectively), whereas mean serum ICTP and urinary NTX values were significantly higher than those of controls (Figs. 4 and 5, respectively). No difference in the values of the biochemical markers was found between male and female patients (data not shown). Mean TBBM was significantly reduced in comparison with normative data (-0.8 ± 0.4 z score,p < 0.001). The degree of reduction in TBBM was significantly greater in male than in female subjects (Fig. 6).

Figure 1
figure 1

Individual values for serum OC concentrations in patients (P) and controls (C). The solid line represents the mean.

Full size image
Figure 2
figure 2

Individual values for serum PICP concentrations in patients (P) and controls (C). The solid line represents the mean.

Full size image
Figure 3
figure 3

Individual values for serum PIIINP concentrations in patients (P) and controls (C). The solid line represents the mean.

Full size image
Figure 4
figure 4

Individual values for serum ICTP concentrations in patients (P) and controls (C). The solid line represents the mean.

Full size image
Figure 5
figure 5

Individual values for urinary excretion of NTX normalized to creatinine in patients (P) and controls (C). The solid line represents the mean.

Full size image
Figure 6
figure 6

Individual values of TBBM expressed as z score in comparison with normative data in male (M) and female(F) patients. The solid line represents the mean.

Full size image

Height, biochemical markers, and TBBM in pubertal patients. Mean height was reduced (-0.8 ± 0.7 z score). Although the degree of reduction in height was greater in male than in female subjects(-1.1 ± 0.6 z score and -0.6 ± 0.6 z score, respectively) the difference was not significant (p = NS). Mean serum OC and PICP concentrations were significantly reduced(Figs. 1 and 2 respectively), whereas ICTP and NTX values were significantly increased in comparison with control subjects(Figs. 4 and 5, respectively). Serum PIIINP levels did not differ from control subjects (Fig. 3). No difference between male and female subjects was found for all the biochemical markers (data not shown). Mean TBBM results were significantly lower than normative data (-1.0 ± 0.4 z score, p < 0.001). Mean TBBM was significantly lower in male than in female subjects (Fig. 6). The values of the biochemical markers and TBBM were not correlated with Tanner's stages within the group (data not shown). Moreover, the values of the biochemical markers and TBBM did not differ between the patients with or without pubertal delay (data not shown).

Height, biochemical markers, and TBBM in young adult patients. Mean height was reduced (-1.0 ± 0.6 z score); the degree of reduction in height was significantly greater in male than in female subjects(-1.3 ± 0.5 z score and -0.6 ± 0.6 z score,p < 0.05, respectively). Mean serum OC concentrations were significantly lower than in controls (Fig. 1), whereas mean ICTP and NTX values were significantly higher than those of controls(Figs. 4 and 5, respectively). Serum PICP levels did not differ from control levels (Fig. 2), whereas serum PIIINP concentrations were significantly higher than those of controls (Fig. 3). The values of the biochemical markers did not differ between male and female patients (data not shown). Mean TBBM was significantly lower in comparison with normative data (-1.1 ± 0.5z score, p < 0.001); a value of TBBM below 2 SD of the normal mean was found in one male subject. Although the degree of reduction in TBBM was greater in male than in female subjects, the difference did not reach significance (Fig. 6). The values of the biochemical markers and TBBM did not differ between the patients with or without pubertal delay (data not shown).

Prevalence of bone fractures in prepubertal, pubertal, and young adult patients, and in control subjects. Six patients (2 pubertal male and 3 pubertal female subjects, and 1 young adult male subject), and 2 control male subjects had suffered bone fractures; prevalence of fractures was 10.2% and 2.8% in patients and control subjects, respectively. In patients, fractures were spontaneous (atraumatic fracture) or as a result of a minor trauma (nontraumatic fracture, arbitrarily defined as a fracture occurring from trauma equal to or less than a fall from a standing height)(41), whereas fractures were caused by a major trauma in the two control male subjects. In patients, upper extremity fractures were the most common (50% of the fractures); one patient had multiple fractures. The fractures healed in all patients without permanent sequelae.

Values of the biochemical markers in patients with or without ultrasonographic liver abnormalities. The values of OC, PICP, PIIINP, ICTP, and NTX did not differ between prepubertal or pubertal patients with ultrasonographic liver abnormalities (prepubertal, n = 8; pubertal,n = 6) and those without (prepubertal, n = 16; pubertal,n = 16) ultrasonographic liver abnormalities (data not shown). On the contrary, young adult patients who showed ultrasonographic liver abnormalities (n = 7) had serum PIIINP levels significantly higher than those without ultrasonographic liver abnormalities (n = 6) (5.7± 0.8 μg/L and 4.0 ± 0.7 μg/L, p < 0.01, respectively), whereas OC, PICP, ICTP, and NTX did not differ (data not shown).

Correlation between height, biochemical markers, or TBBM and disease severity. Height z score, ICTP and NTX values, and TBBM z score were significantly correlated with disease severity, expressed as the Shwachman-Kulczycki score (r = 0.36, p< 0.001; r = 0.27, p < 0.04; r = 0.41,p < 0.001; and r = 0.29, p < 0.04, respectively), or Chrispin-Norman score (r = -0.41, p < 0.001; r = -0.34, p < 0.02; r = -0.45,p < 0.001; and r = -0.38, p < 0.01, respectively), whereas OC, PICP, and PIIINP did not (data not shown). The values of the Shwachman-Kulczycki score were significantly correlated with those of Chrispin-Norman score (r = -0.54, p < 0.001).

Height z score was significantly correlated with TBBM z score (r = 0.47, p < 0.001), whereas no relation between the biochemical markers and height or TBBM z scores was found (data not shown). BMI significantly correlated with PICP and ICTP levels(r = 0.36, p < 0.005 and r = 0.25,p < 0.05, respectively) or TBBM (r = 0.55, p< 0.005), but not with OC and PIIINP (data not shown).

DISCUSSION

The increased life expectancy of patients with CF resulting from the new strategies of treatment(42) has increased the prevalence of disturbances in bone mineralization and growth. In prepubertal patients we demonstrated augmented ICTP and NTX values that reflected an increased bone resorption rate. Likely, the effect of somatic growth on ICTP and NTX values was poor as the linear growth of these patients was slightly decreased. In pubertal patients, reduced OC and PICP levels suggested a decreased bone formation rate, whereas increased ICTP and NTX concentrations reflected an augmented bone resorption rate. These data indicated that pubertal patients had an impaired bone turnover. However, the diminished somatic growth may contribute to reduce their OC and PICP concentrations. This effect appeared less clear for ICTP and NTX. On the whole these findings seem to indicate that ICTP and NTX were more markers of precocious impaired bone turnover than were OC and PICP. In young adult patients, reduced OC and increased ICTP and NTX values certainly reflected an impaired bone turnover, because linear growth had ceased in these patients.

Impaired bone turnover likely was a main cause in reducing TBBM in our patients. The major degree of reduction in TBBM in young adult patients in comparison with prepubertal and pubertal patients suggested that the disease severity progressively affected bone mineral status, as indicated by the relationship between TBBM and the Shwachman-Kulczycki score or Chrispin-Norman score. Indeed, a positive correlation between disease severity and regional(3, 5, 9) or whole body bone mineral density(9) was previously also reported. On the contrary, disease severity was not predictive of bone mineral status in the study of Bachrach et al.(7). In our study the reduction in bone mass was greater in male than in female subjects, as previously observed by Gibbens et al.(3) and Grey et al.(5), but in contrast to these authors(3, 5) we did not find a major degree of disease severity in our male patients. Mischler et al.(1) found that patients at the greatest risk for bone demineralization were adolescent girls, whereas Bachrach et al.(7) did not show any difference in bone mineralization between male and female adult patients. These results suggested that, in addition to disease severity, other factors as treatment regimens or the genetic structure might influence bone mineral status in CF patients. The values of the biochemical markers and TBBM were similar in patients with pubertal delay and in those with a normal tempo of puberty, suggesting that the delayed sexual development was not the main cause of the reduced bone mineralization in our CF patients. At any rate, the small number of patients for each group precludes definitive conclusions.

The increased fracture rate we found in our pubertal CF patients confirmed the results of Henderson and Specter(8). On the contrary, none of the patients of Grey et al.(5) had suffered fractures, even though they showed a reduced bone mineral density. It has been suggested that fracture risk is already significantly increased if bone mineral density falls below 1 SD of normal mean(43). Although only the young adult patients had a mean TBBM value below this limit, fracture rate was higher in pubertal patients than in young adult patients. These data suggested that the degree of reduction in TBBM contributed only in part to the increased fracture rate in CF patients. Therefore, an impaired bone quality or an inadequate accumulation of bone mass could be adjunctive factors increasing the fracture rate in pubertal CF patients.

Some studies in asthmatic patients have demonstrated that inhaled steroids may affect bone turnover and collagen metabolism(44–46) and bone mineralization(47). However, we did not find any difference in the values of the biochemical markers and in TBBM between patients receiving beclomethasone dipropionate and those who did not receive this treatment, suggesting that the reduction in the biochemical markers and TBBM was likely not due to the effect of the inhaled steroid.

Fibrosis consists primarily of collagen fibrils(13) and may be directly documented by invasive methods such as biopsy. The quantification of fibrosis by biopsy is not entirely reliable, given the small size of the sample and the heterogeneous distribution of fibrosis(48). Moreover, iterative biopsies are ethically questionable, mainly in children, and dubiously reliable for monitoring the progression of fibrosis. The elevated PIIINP levels we found in young adult patients suggested an increased activity of the fibrotic process, as linear growth had ceased. Indeed, serum PIIINP has been claimed to be a means of differentiating fibrotic and nonfibrotic chronic liver diseases(49). In addition, it has been shown that PIIINP levels correlated with the activity of the fibrotic process(17). Young adult patients with liver ultrasonographic abnormalities had significantly higher PIIINP levels than those without liver ultrasonographic abnormalities; however, liver ultrasound examination as well as changes in liver enzymes are not sufficiently sensitive or specific for the detection of liver fibrosis or cirrhosis(50). Pulmonary fibrosis(18, 19) and diabetic microangiopathy(20, 21) in the two patients who developed insulin-dependent diabetes mellitus likely contributed to increased circulating PIIINP levels. Therefore, measurement of PIIINP may be a noninvasive marker for the detection of precocious development of the fibrotic process in CF patients. With regard to PICP as an index of fibrosis, the normal values we found in young adult patients may be related to the fact that, during a fibro-proliferative response, the deposition of type III collagen precedes that of type I collagen; whereas in advanced fibrosis, type I collagen predominates(51). However, the predominance of type I collagen in late fibrosis can also be due to its slower and more incomplete turnover. Moreover, the evidence that PICP was not elevated similarly as PIIINP could be explained by the fact that the majority of circulating PICP originates in the skeleton so that the reference interval of PICP is relatively wide, and the small increase caused by the liver may not be evident. Indeed, at least during wound healing, both PICP and PIIINP increased simultaneously and to the same extent(52). The normal PIIINP values we found in prepubertal and in pubertal patients may suggest that the fibrotic process was probably poor in these patients. However, it must be considered that the minor elevation of PIIINP levels in prepubertal and pubertal patients in comparison with young adult patients may also be related to the higher PIIINP levels that physiologically occur during linear growth as suggested by Trivedi et al.(16). Therefore, measurement of PIIINP levels may have a limited value in estimating the fibrotic process during childhood.

In conclusion, our study demonstrated that CF patients have impaired bone turnover and collagen metabolism, and reduced bone mineral status. The increased values of ICTP and NTX in prepubertal, pubertal, and young adult patients suggested that an increased bone resorption rate was a main factor in impairing bone turnover. The reduced values of OC and PICP in pubertal patients, or OC in young adult patients, suggested that reduced bone formation rate was an adjunctive factor affecting bone turnover. The reduction in TBBM probably reflected the impaired bone turnover. The study also showed that measurement of PIIINP may be an useful index to detect the progress of fibrotic process in young adult patients; whereas it seems to have a limited value in prepubertal and pubertal patients. Although the biochemical markers we measured were unable to independently examine the effect of the disease on bone turnover, collagen metabolism, and somatic growth, they may represent a useful tool in the management of these patients. Indeed, these markers can be measured widely and repeated several times in a single patient, and their measurement is not invasive. At any rate, further studies are needed to establish the usefulness of these biochemical markers to optimize the strategy of treatment and to identify their prognostic value.