INTRODUCTION

Cystic fibrosis (CF) is an inherited multisystemic disease caused by the alteration of a gene located on chromosome 7 (CFTR: cystic fibrosis transmembrane conductance regulator) (¹). Alterations in CFTR lead to an impaired transport of chloride, bicarbonate, and sodium ions across epithelial cell membranes (²). In healthy subjects, mucus is mainly composed of mucins and water. Hydration and pH regulate mucus viscosity, and both functions are controlled by CFTR on the apical surface of epithelial cells. Chloride movement dictates the degree to which mucus retains water, while CFTR-mediated bicarbonate efflux plays a key role in defining pH, which is critical for a healthy antibacterial response. When the CFTR protein is altered, thick mucus secretions are produced (¹). These thickened secretions are produced predominantly in the pancreas and lungs. This disturbance at the pulmonary level may cause decreased pulmonary function. In addition, malabsorption and maldigestion of nutrients, particularly fat-soluble vitamins and fats, can lead to poor nutritional status. Until a few years ago, malnutrition was considered to be associated with CF because it was practically always present at the time of diagnosis and a vast majority of patients suffered from deterioration of their nutritional status and died deeply malnourished. The interaction between lung function and nutrition has a great importance, as a parallel worsening in both would affect prognosis and quality of life. Thus, malnutrition behaves as a risk factor predictor of morbidity and mortality in CF (¹-⁶). Although the prevalence of malnutrition has decreased considerably, figures close to 25 % continue to be reported in both children and adults (⁷).

Guidelines on endocrine-nutrition care for patients with CF at different stages of life (infants, children, and adults) have been developed by the ESPEN, ESPGHAN, and ECFS. These guidelines recommend periodic nutritional assessments as a primary step to achieve the best therapeutic and prognostic outcomes as possible (⁴). At every visit, patient's weight, height, body mass index (BMI), and weight loss over time are recorded (⁸). BMI is used in the clinical setting to quantify the nutritional status of CF patients. BMI targets are used in clinical practice; however, although the relationship between BMI and lung function in CF patients is well established, the exclusive use of BMI as an indicator of nutritional status can be misleading (⁹). In clinical practice, individuals with the same BMI may have different distributions of fat (FM) and fat-free mass (FFM); besides, patients with normal BMI may lack of FFM. Measuring the proportion of FFM and lean body mass may help to describe better the nutritional status in CF. In CF, one of the most important determinants of morbidity and mortality related to malnutrition is the decrease in fat-free mass (FFM); therefore, it is important to measure it accurately (¹⁰). Different methods can be used to assess whole-body and segmental BC: dual energy X-ray densitometry (DXA), bioimpedanciometry (BIA), skinfold measurement (SM), or deuterium dilution (DD). In addition, imaging methods such as computed tomography (CT) or magnetic resonance imaging (MRI) can also be used for evaluating nutritional status. However, there has been great variability among the studies that assess body composition (BC) by different methods in CF patients (¹¹). Therefore, clear conclusions to propose an evidence-based algorithm to assist in the assessment of BC cannot be drawn. Currently, DXA represents the preferred option for assessing and monitoring BC changes in this population, and is the most frequently applied method (¹⁰). Furthermore, in clinical practice DXA is considered the technique of choice (¹²). Thus, it is necessary to know the degree of agreement between the values obtained by DXA and other techniques more easily implemented in routine care, such as SM, which is more accessible for health care teams. In addition, increasingly accurate equations are available for the assessment of BC, which would support the use of SM (¹³). However, few studies have assessed BC using different techniques and evaluated their agreement in adults with CF.

We hypothesize that SM has both an adequate correlation and concordance for measuring FFM with respect to DXA in adults with CF.

The objective of our study was to describe BC using different techniques to evaluate nutritional status (DXA vs SM), and to assess their concordance in adults with CF.

MATERIALS AND METHODS

This was a cross-sectional, observational study in patients with CF in situation of clinical stability. A sequential recruitment was carried out when patients attended the CF adult's outpatient consulting room for their annual examination.

RESULTS

Forty-one patients with CF were recruited. Of these patients, nineteen were males (46.3 %) and twenty-two (53.7 %) were females. Their median age was 29.8 (20.9-33.7) (Table I).

Table I.  General characteristics 

BMI: body mass index; FEV1: forced expiratory volume in 1 second; FVC: forced vital capacity; DXA: dual-energy X-ray absorptiometry; FM: fat mass; FFM: fat-free mass; FFMI: fat-free mass index; SM: skinfold measurement.

Only four (9.8 %) patients had a BMI lower than 18.5 kg/m². By DXA, FFM (kg) results were median: 52.8 (interquartile range, 47.8-56.9) kg with FFMI 17.9 (16.7-19.3) kg/m² in men, and 36.7 (33.1-38.9) kg in women with FFMI 14.7 (14.2-15.8) kg/m². Using this technique, twenty (48.6 %) patients presented FFM malnutrition, with 16.7 % of males and 59.1 % of females being affected. By SM, FFMI was 18.7 (17.2-20.0) kg/m² in males and 14.9 (14.2-15.8) kg/m² in females; moreover, sixteen (39.1 %) patients presented FFM malnutrition, with 20.8 % of males and 61.8 % of females being affected. The general characteristics of the patients are summarized in table I. All patients designated as malnourished by FFM according to SM were found to be malnourished according to DXA.

Figures 1A and B show body composition as assessed by DXA and SM. They show that the prevalence of FFM malnutrition is higher for DXA than for SM. In addition, the prevalence of malnutrition was higher in females than in males in the nutritional assessment by both techniques.

Figure 1 A. Body composition according to sex by DXA (%). B. Body composition according to sex by SM (%). 

Concordance was high between SM and DXA for the FFM in kg (ICC of 0.950; p < 0.001) and also in % (Table II). Concordance was also high for FM as expressed both in kg and %. When the techniques were compared by applying the ESPEN criteria for FFMI malnutrition, concordance was moderate: kappa coefficient of 0.440 when comparing SM with DXA (p = 0.006).

Table II.  Comparison of measurements and agreement between BC values by SM and DXA 

Medians. IQR: interquartile range; ICC: intraclass correlation coeffi cient; DXA: dual-energy X-ray absorptiometry; SM: skinfold measurement; FM: fat mass; FFM: fat-free mass. p1: statistical signifi cance of the comparison of DXA and SM measures; p2: statistical signifi cance of CCI.

By analyzing individual differences using the Bland-Altman plot analysis, the degree of agreement between baseline BC data obtained with DXA and SM was assessed (Fig. 2). The mean overestimation with respect to DXA was +1.44 ± 0,62 kg, with a tendency to greater dispersion in higher FFM.

Figure 2.  Comparison between FFM by SM and DXA. 

DISCUSSION

In our study we found a high prevalence of FFM malnutrition, especially in females, according to the different nutritional assessment techniques used. In addition, we found a high statistical agreement between DXA and SM for FM and FFM values. However, the concordance when we assessed malnutrition according to FFMI was moderate. It was observed that SM tends to overestimate FFM and to underestimate FM, compared to DXA.

According to the Cystic Fibrosis Foundation (CFF), a BMI ≥ 22 kg/m² in females and a BMI ≥ 23 kg/m² in males is recommended as a target for adults, as it is associated with improved lung function (²³). However, despite the improvement in disease progression and nutritional status, according to the latest European Cystic Fibrosis Registry the prevalence of malnutrition determined by BMI (below 18.5 kg/m²) is approximately 30 % of adults (²³). In our study, the mean BMI is slightly higher than the one published in the 2014 European CF registry (22.2 vs 21 kg/m²) for a similar mean age (²⁴). In addition, the rate of malnutrition determined by BMI in our CF population was 9.8 %, lower than the one published in other series (²⁴,²⁵). However, BMI measurement alone in adults with CF should not be the only parameter used to assess nutritional status. Engelen et al. have warned about the increased prevalence of overweight and obesity in this population (even in association with severe mutations), and BMI may therefore be a confounding factor for nutritional status. It has been observed that between 25-38 % of adult CF patients with a normal or high BMI have low lean mass levels (²⁶). Thus, guidelines on nutrition care for patients with CF recommend evaluating BC as part of the nutritional assessment (³,⁴).

In CF, one of the most important determinants of morbidity and mortality in relation to malnutrition is the decrease in FFM (¹⁰,²⁷). It is therefore useful to measure BC, using different techniques, to estimate FFM; in this sense, the calculation of FFMI is particularly useful (⁴). In the review by Calella et al. they conclude that the results showed great variability in the methods used to evaluate BC in patients with CF (¹⁰). In the review by Gomes et al. they conclude that further studies are needed to identify and determine which FFM measurements are associated with improved lung function and nutritional status in CF patients (⁵). The use of DXA-derived FFM measurements in clinical practice among CF patients shows great potential utility and value. Furthermore, together, FFM and BMI may provide a more comprehensive picture of nutritional status during the nutritional evaluation of CF patients (⁵).

In the present study, we have performed a comparison of simpler techniques as SM with DXA. Some of the potential advantages of DXA for assessing BC are the high accuracy of SM and DXA (²⁸), the good agreement with the results obtained using the 4-compartment model (²⁹), good short-term reproducibility (³⁰), and the possibility of assessing total and segmental BC. On the other hand, the limitations of DXA are the lack of portability of the equipment, the radiation exposure associated with the scan, or the fact that it requires the patient to be in a decubitus position for a few minutes (which may be difficult for patients with severe lung disease) (²⁸). For this reason, several studies have questioned whether DXA could be replaced by other measures of BC (³¹). A recent study of children with CF found, using DXA, estimated a prevalence of malnutrition of 38.6% based on FFM deficit (defined as FFMI < 10^th percentile) (²⁷). In our work, the median and IQR for FFMI determined by DXA were 17.9 (16.7-19.3) kg/m² in males and 14.7 (14.2-15.8) kg/m² in females, being slightly higher when determined by SM. By DXA, we found a prevalence of FFM malnutrition of 48.6 %, which was much higher in females (68.4 %) compared to males (27.8 %) and using the criteria recommended by ESPEN (17 kg/m² in males and 15 kg/m² in females) (⁸). It is possible that the ESPEN criterion for females is too high in this population, overestimating the prevalence of FFM malnutrition. According to the Spanish population data, the FFMI cut-off point for males would remain at 17 kg/m²; however, in females it would be 14.4 kg/m² (³²,³³). If we consider 14.4 kg/m² as the cut-off point for FFMI in females to diagnose FFM malnutrition, the prevalence of malnutrition in females in our study would drop to 47.4 % using DXA.

Assuming that subcutaneous adipose tissue represents a constant proportion of total body fat, and that measurement sites are representative of the average subcutaneous adipose tissue thickness, skinfold thickness measurements are useful to estimate FFM and FM (²⁹). This may vary depending on race, age, gender, and disease (²⁹). Inter- and intra-observer variability are other limitations of this technique (³⁰). The use of SM in CF has been evaluated in several studies (²⁷,³⁰,³⁴-⁴⁰). The same researcher always performed the anthropometric measurements in our study. We observed that 39.1 % of patients had FFM malnutrition, with 16.7 % of males and 59.1 % of females being affected, estimated by SM. This prevalence is slightly lower, especially in females, which could classify malnourished people as normonourished. In this sense, if we used the cut-off point 14.4 kg/m² (³²,³³) to detect malnutrition according to FFMI, the prevalence of FFM malnutrition would decrease considerably to 13.6 % according to SM.

Chomtho et al. evaluated the use of upper arm anthropometry compared to DXA in healthy children and children with CF, concluding that upper arm anthropometry is less accurate than DXA for determining segmental and total FFM. The results were better for FM determination (³⁵). In the work by Alicandro et al. using DXA as a reference method, they conclude that BC estimation obtained by SM or BIA cannot be part of the standard nutritional assessment of CF patients due to low precision at the individual level, at least until reliable CF-specific equations become available (³⁶). King et al. also compared BIA and SM with DXA in adults with CF (³⁷). SM overestimated lean mass by almost 2.4 kg on average compared to DXA. In a study of our group in people with non-CF BQ, SM also overestimated lean mass with very similar values (2.35 kg on average) and was homogeneously distributed across lean mass values (³⁸). The study by de Meer et al. that assessed changes in lean mass by skinfold measurement in malnourished children with CF after a physical exercise program concluded that regardless of disease severity this technique is applicable to detect changes in lean mass (³⁹). Stettler et al. also assessed prospectively body composition in children with CF using various techniques such as double watermarking, SM and BIA; the study showed high concordance for lean mass but not for FM (⁴⁰). In our study, there was good statistical agreement between SM and DXA for assessing FM and FFM. However, the concordance was moderate for detecting malnutrition according to FFMI. Using Bland-Altman plot analysis, the mean overestimation with respect to DXA was +1.44 ± 0,62 kg, thus, the degree of agreement showed a greater dispersion as FFM increased.

The limitations of our study are mainly due to the fact that it is a single-center study, which prevented us from having a larger sample size. Furthermore, the study design was cross-sectional. These limitations hamper drawing causal conclusions, and therefore, we can only speculate about different associations.

In conclusion, the prevalence of FFM malnutrition is high in CF patients, despite presenting a normal BMI, especially in females. The cut-off point for FFM proposed by the ESPEN may be high in our population. There was good statistical agreement between SM and DXA for assessing FM and FFM. However, the concordance was moderate for detecting malnutrition according to FFMI, and the degree of agreement showed a greater dispersion as FFM increased. Therefore, DXA remains the technique of choice for nutritional assessment in CF, and SM can be used in cases in which it is not available.