The Prognosis of Cystic Fibrosis – A Clinician’s Perspective

Looking at the prognosis of Cystic Fibrosis (CF) from the clinician's point of view is very relevant. Median predicted survival age of CF increased from 6 months when the disease was first described (1938) to 12 years in 1970 and over 35 years in 2010 in the United States of America (Davis, 2006). Three types of factors weigh on this prognosis, which is conditioned by lung disease: factors linked to the quality of care management, to genetics and to the environment. It has been estimated from studies in twins and siblings that the relative influence of the latter two is roughly equivalent. Though pollution may have increased in certain areas, and lighter forms of CF are now being detected by neonatal screening and in clinics where nasal potential measurements are widely available, their impact is limited and can not account for the spectacular changes in life expectancy. Thus, this improvement is due essentially to a better care management. Quality of care is the main determinant of CF prognosis.


Introduction
Looking at the prognosis of Cystic Fibrosis (CF) from the clinician's point of view is very relevant. Median predicted survival age of CF increased from 6 months when the disease was first described (1938) to 12 years in 1970 and over 35 years in 2010 in the United States of America (Davis, 2006). Three types of factors weigh on this prognosis, which is conditioned by lung disease: factors linked to the quality of care management, to genetics and to the environment. It has been estimated from studies in twins and siblings that the relative influence of the latter two is roughly equivalent. Though pollution may have increased in certain areas, and lighter forms of CF are now being detected by neonatal screening and in clinics where nasal potential measurements are widely available, their impact is limited and can not account for the spectacular changes in life expectancy. Thus, this improvement is due essentially to a better care management. Quality of care is the main determinant of CF prognosis.
The question of prognosis is almost invariably the very first that parents of a newly diagnosed CF infant will ask their physician, whose task to answer in a sensitive and sensible fashion is by no means easy. This question is also at the heart of daily concerns of the clinician, who has to take its determinants into consideration not so much in view of their fascinating underlying mechanisms (e.g. modifier genes) but rather to the extent they can give grip to improved care.

4
Limiting the analysis to patients homozygous for the F508del mutation could be even 'cleaner ' (Zelin et al., 2010;Lebecque et al., 2010). It is important to note that over the past 60 years, median predicted survival has actually increased in a continuous fashion by almost 6 years every decade (Davis, 2006). Part of this improvement is probably linked to the increased detection of milder forms of the disease. As proposed for the comparison of registries, it would be interesting to study the data of only those patients homozygous for the F508del mutation. For four reasons developed below, this brutal landmark has limited value in an individual patient.

Cohort effects
Cohort survival curves consistently show that survival continues to improve with each successive birth cohort over the decades. However, this effect is not taken into account by current survival curves. As a result, advice based on the latter could be unduly pessimistic. This has led authors to model the trend observed in cohort survival curves and to extrapolate a median survival for recent cohorts. Accordingly, a median life expectancy of the order of 40 years was predicted for newborns in 1990 in the UK (Elborn et al., 1991). This proved to be realistic and updated extrapolation for the birth cohort of the year 2000 predicts a median survival of 50 years (Dodge et al., 2007). Recent work has further validated this approach (Jackson et al., 2011).

Wide heterogeneity of the disease
Median predicted survival does not take into account the vast heterogeneity of CF. Yet the latter has long been recognized, even amongst patients homozygous for the F508del mutation Johanssen et al., 1991). When diagnosing a new patient carrying 2 CF-causing mutations, this variability renders precise individual prognosis almost always impossible.

Quality of life
The raw median predicted survival rate says nothing about quality of life. In CF, pulmonary function is often used as a surrogate for survival, with FEV 1 remaining the single most useful parameter (Kerem et al., 1992). Though insensitive to early stages of the disease, spirometry is widely available, inexpensive, non-invasive and very reproducible. It can usually be performed from the age of 5 and upwards. FEV 1 has the advantage of reflecting pulmonary involvement, thereby conditioning prognosis, throughout the whole course of the disease. The rate of FEV 1 decline might be an even stronger surrogate for survival (Liou et al., 2001;Schluchter et al., 2002;Rosenbluth et al., 2004).
Quality of life is at least as important as its length. For every human being, it is largely conditioned by how an individual handles the careful balance of renunciations, and accepts these. Its precise and fine perception inevitably escapes all questionnaires. Specific tools developed over the past 20 years cannot presume to its assessment but can help discern the impact of new treatment modalities and are increasingly being used in this context (Abbott et al., 2011). This also implies that, given the choice between equally efficient treatments, the least invasive treatment and follow-up modalities are to be favoured (Wainwright et al., 2011). Under close supervision and for adequately selected patients, home intravenous

Prognostic factors
A pragmatic way of looking at prognostic factors of CF is to distinguish those that are linked to the quality of care management and those on which care management has little (e.g. environmental factors) or no grip (genetic factors) (Wolfenden et al., 2009). Figure 1 gives an overview of prognostic factors in CF.
Three aspects of the discussion are worth mentioning i) interactions between factors are numerous and their complex statistical evaluation study has probably been too shallow in many publications; ii) the evaluation of the role of certain less easily quantifiable factors, such as treatment adherence, escapes usual means of analysis; iii) our choice is to focus on factors that appear essential to a clinician. The complexity of interactions between the different factors is clearly illustrated by the example of the impact of passive smoking on respiratory function. Although the deleterious effects of passive smoking are well established, they vary according to CFTR genotype and certain alleles of modifying genes (Collaco et al., 2008). Moreover, several studies demonstrated a link between passive smoking and socio-economic status (Smyth et al., 1994).

6
One example of 'the clinician's choice' in the discussion is not to dwell on the often reported issue of a 'female disadvantage' (Rosenfeld et al., 1997;Mehta et al., 2010;Olesen et al., 2010), which may not concern adult diagnosed CF patients (Nick et al., 2010). Numerous hypotheses have been brought forward to account for it, without any practical impact to date. Furthermore, though this might have been achieved at the price of a higher burden of treatment in females (Olesen et al., 2010), there is evidence that modern intensive treatment may result in similar key clinical parameters for the two genders (Verma et al., 2008;Olesen et al., 2010).

Good care management of CF
Early and mostly optimal management of CF is thus the principal reason for the improved prognosis observed over the past decades. Comprehensive follow-up and progress in symptomatic treatment proved essential. Antibiotics are the mainstay of CF therapy. In view of the complexity of this multisystemic and in many ways unique pathology, optimal management is only possible in truly specialized structures. Limiting the impact of economic status on treatment is an important issue. Most impressively, a pioneer clinician had already specifically stressed the importance of every single one of the above-named factors as early as 1974 (Crozier, 1974).
Taken in isolation and often combined to other factors (see discussion below), several complications can be linked to an accelerated rate of FEV 1 decline. Their awareness, prevention or early detection followed by optimal management carries the potential to 7 reduce the rate of FEV 1 decline. This is well illustrated by the case for CF-related diabetes (CFRD) over the past 20 years. At one large CF centre, early detection and optimal management of CFRD resulted in a decrease on its impact on mortality and the disappearance of a sex difference in mortality (Moran et al., 2009).
Two factors in particular, chronic airway colonisation by Pseudomonas aeruginosa (PA) and malnutrition, have such spontaneous prevalence and such prognostic impact, that their prevention constitutes one of the major objectives of CF care management.
Several medications have been associated to slowing down the rate of FEV 1 decline: dornase alpha (Konstan et al., 2011), ibuprofen (Konstan et al., 1995;, azithromycin (Hansen et al., 2005) and inhaled corticosteroids (De Boeck et al., 2011). This does not mean that all patients will benefit from them. In addition, possible side-effects in the long-term have to be kept in mind (cf. e.g. ibuprofen). Also, it cannot be inferred that other drugs are necessarily ineffective: they may simply not have been subjected to appropriate studies, which occasionally may be due to commercial reasons or ethical concerns (e.g. physiotherapy).

Early diagnosis: CF newborn screening (NBS)
A number of studies using bronchoalveolar lavage (BAL) markers of infection and inflammation, lung function tests or computed tomography of the chest have documented that significant lung damage occurs very early in many, even asymptomatic, infants (Khan et al., 1995;Ranganathan et al., 2001;Davis et al., 2007;Mott et al., 2009). The results of two recent studies concerning very young infants who were investigated routinely after CF diagnosis by NBS are particularly striking (Sly et al., 2009;Stafler et al., 2011). Unsuspected positive cultures were found in 21-27% of them, there was evidence of airways inflammation with BAL neutrophilia in most patients and CT evidence of bronchial dilatation in 18.6%.
CF is a progressive disease for which symptomatic treatment has proven to have a real impact albeit of partial efficiency. So, starting treatment as early as possible is meaningful. CF NBS is now available throughout the USA and in many European countries. For those newborns carrying a genotype clearly associated with the disease, benefits of neonatal screening has long been proven in nutritional terms. Respiratory benefits have only recently been acknowledged (Accurso et al., 2005;Rosenfeld et al., 2010), as these were temporarily obscured by a publication indicating an increased risk of early chronic colonisation by PA in screened newborns (Farrell et al., 2003). This increased risk was eventually linked to the lack of measures aiming to limit cross infections in that particular centre and subsequent studies failed to confirm it (Siret et al., 2003;Sims et al., 2005;Baussano et al., 2006;Collins et al., 2008). In countries with a high level of medical care, neonatal screening enables clinicians to diagnose CF in infants before the age of 2 months, with demonstrable benefits (Sims et al., 2007).

CF centres are necessary
CF NBS is universally coupled with immediate referral to a specialist centre. Guidelines and international consensus all emphasize this to be the key to efficiency for all CF programs www.intechopen.com Cystic Fibrosis -Renewed Hopes Through Research 8 (Castellani et al., 2009;Comeau et al., 2007). Even outside the context of CF NBS, early referral to a specialist healthcare centre is considered as a major prognostic factor as highlighted in consensus reports on optimal management of CF (Littlewood, 2000;Kerem et al., 2005;Colombo et al., 2011). Though this has long been supported by a number of common sense reasons and has been mentioned as a model for the management of complex diseases (Schechter et al., 2005), most published studies concerning the impact of this centralization on lung disease are either biased due to comparison with historical controls, and/or probably underpowered (Hill et al., 1985;Nielsen et al., 1988;Walters et al., 1994;Collins et al., 1999;Merelle et al., 2001;Van Koolwijk et al., 2002).
Two studies avoid these pitfalls. Mahadeva et al. compared two groups of adults who had either received continuous care from paediatric and adult CF centres (n=50) or had received neither paediatric nor adult centre care for their CF (n=36). Excluding body mass index as a covariate, FEV 1 was significantly better in the first group (Mahadeva et al., 2000). More recently, a Belgian retrospective multicentre study clearly showed that earlier referral of children suffering from CF to specialist care was associated with significant pulmonary benefits (Lebecque et al. , 2009). Children referred 'early' (less than 2 years after diagnosis) had a better FEV 1 (86.7% pred. ± 19.4 vs. 77.2% ± 22.4, p=0.01) and a lower prevalence of PA (17.5% vs. 38.6 %, p<0.05) than carefully matched patients referred later.

CF centres might not be sufficient
Large differences between outcome variables obtained from the different centres have been recognized for a long time (Bauernfeind et al., 1996) and are now drawing considerable attention as they may provide an opportunity to develop quality improvement initiatives. Striking illustrations of this outcome heterogeneity are provided by recent data from the 2007 US Registry (CFF Registry, 2007 -public data), allowing for comparisons between centres: i) mean FEV 1 (% predicted) for CF children aged 6-17 years ranged from 75 to 103% (national average: 92.6%, reference values: Wang & Hankinson) ii) the percentage of patients under 20 years of age with a BMI < 5th percentile ranged from 32 to 83% (national average: 52.7%) iii) M R S A in f e c t ion r a t e ra ng e d f r om 6 to 42% (national average: 21.2%). These differences persisted after taking into account socio-economic factors. Similar features can be derived from the CF German Registry where a quality management program with an overall coverage of 82% for the year of 2005 confirmed considerable differences between centres in terms of key parameters (Stern et al., 2008). For instance, the percentage of children (6-18 years) with an FEV 1 above 80% of the predicted value ranged from 20 to 100% in centres treating less than 50 patients and from 35% to 100% in larger centres. Globally, the mean FEV 1 in this age group was 88% of the predicted value. A recent Belgian multicentre study confirmed the broad outcome differences between reference centres within this small country, where corresponding values for FEV 1 in children reportedly ranged from 74% to 95% predicted (Lebecque et al., 2009) while the prevalence of PA (last visit of the year) ranged from 5 to 46% and the mean weight expressed as percentage of ideal body weight ranged from 88% to 100%.

Marked differences in clinical results between CF centres: Why?
One report of 18,411 patients followed in 194 North American centres in 1995 showed that close monitoring and heavier treatment (more frequent antibiotics in particular) clearly characterized those north-American centres with better clinical results (Johnson et al., 2003). It is interesting to note that this more intensive management included two modalities which do not correspond to standard management: frequent prescription of nedocromil and more common prophylactic prescription of inhaled antibiotics, particularly in patients with still little pulmonary involvement. The use of nedocromil in CF is not well documented in the literature, neither is the potential of giving prophylactic inhaled antibiotics, which is discussed a little further in this chapter. Another North American study concerned 837 children aged 6 to 12 years (Padman et al., 2007), whose CF centres were classified according to the children's FEV 1 values in 2003. The analysis also suggested that closer follow-up starting before the age of 3 years was characteristic of those centres with the best functional results. This 'more attentive and more active 'attitude' is in line with current standards of care, which are derived as much as possible from evidence-based medicine (Littlewood, 2000;Kerem et al., 2005;Littlewood, 2005;Tiddens, 2009).
In 2002, the CFF launched an innovative Quality Improvement (QI) initiative, which included the vision statement that within 5 years, life expectancy of CF patients could be extended by 5-10 years through the consistent implementation of existing evidence-based clinical care (Quinton, 2007). However, evidence-based medicine has its limitations and possible stumbling blocks (Driever, 2002;Saarni et al., 2004;Miles et al., 2008;Shahar, 2008;Miles et al., 2011), in particular when it concerns diseases as complex as CF for which numerous therapeutic modalities have simply not yet been adequately studied (Cheng et al., 2000;David, 2001;Briggs et al., 2006;Kraynack et al., 2009). Best practice remains individualized in part, although a coherent foundation of evidence-based medicine is absolutely necessary.
One comparative study of 3 CF centres illustrates how even beside care standards, clinician intuition can lead to potentially judicious choices that deserve prospective studies. A better nutritional status was indeed linked to the prescription of dornase alpha (Pulmozyme ®) under the age of 2 years (Padman et al., 2008), whereas the effects of Pulmozyme in this age group have only been reported in one limited study (Berge et al., 2003) where 9 infants were given the drug for 2 weeks. The necessity to conceive and study new strategies stems from the fact that current ones are relatively powerless in preventing pulmonary lesions in early diagnosed infants (Stick et al., 2009).
Information concerning the standards of CF care is widely available as are new drugs in rich countries. Differences in clinical results between various CF centres must therefore have a different origin. One can suspect subtle combinations of several factors resulting in small differences in interventionist attitudes. Beside adequate means i.e. infrastructures and human resources, optimal care requires a careful organisation (structure of the centre, coherence of attitudes, multiple fail safe systems etc.) necessitating considerable long-term thoughtful investment of all those involved.

Variability of outcomes in CF centres: Benchmarking as a right to patients
It is not possible for a CF clinic to function without landmarks. Specific databases prove to be very precious by giving immediate access to progress over time and by age group of essential parameters, thereby constituting a sort of 'compass' for measures that could be taken to improve care management. Although limiting the objective to FEV 1 improvement would be reductive, it clearly remains an essential parameter. The FEV 1 of patients aged 6-18 years old ( Figure 2) is a particularly relevant landmark for the 3 following reasons: i) very few patients die or require transplantation prior to that age ii) a significant proportion of adults has not benefited from immediate specialized care iii) adolescence is a particularly high risk time period with regard to deterioration of respiratory function. Appropriate indicators to define nutritional outcomes are also necessary (Lai et al., 2008). If they are to be used to compare CF centres, these indicators have to be adjusted for risk factors -which does not change the overall outcome variability of CF centres (Schechter et al., 2002) -and several years should be taken into consideration, in order to appreciate their coherence and tendencies.  (Wang et al., 1993;Hankinson et al., 1999).
The differences in mean FEV 1 observed in children from various CF centres can exceed 20% of predicted values. Recent data can help to perceive the implications of such differences: i) some centres now report a mean annual FEV 1 decline around 1% (Que, 2006) ii) in terms of FEV 1 (expressed in % pred.), improvements from baseline observed in two major randomized controlled trials of inhaled dornase alfa and tobramycin were less than 4% and 5% respectively (Fuchs et al., 1994;Ramsey et al., 1999) iii) in the USA, from 1990 to 2008, median FEV 1 at 12 years of age increased by 14 % predicted. Thus the prognostic significance of the differences between centres is major, also in terms of life expectancy. For this reason, public access to these data can be considered as a right to the patient and/or parents. In 2006, the outcomes data from the CF Foundation registry became public on the foundation's website, with an accompanying warning against simplistic interpretations. This choice bore witness to the priority given to the patients. Their health is at stake, and keeping them informed has to take precedence over the risk of embarrassing some physicians. In the USA, the fear that some patients may leave CF centres with average performance has not materialised: patients and their family maintain their trust once they realize that everything is done to improve results (Quinton, 2007).

'Good enough' care for CF does not exist: Towards a culture of improvement
Benchmarking, which is the process of identifying practices associated with the best results, may help to understand why certain centres obtain better clinical results. Not only can it be considered a right to the patients, it can also serve as an incentive to CF centres to increase their efforts to improve quality of care delivery. In the USA, most CF healthcare providers actually accepted it as a call to action (Schechter et al., 2005). Public release of meaningful adjusted clinical outcomes data has been used in other fields of medicine, such as cardiac surgery, as a means to stimulate improvement efforts within the medical profession (Ferris et al., 2010). Others have advocated 'softer' uses of benchmarking which do not require to be channelled via public data (Stern et al., 2011). In the field of CF, benchmarking is not an easy task. It is possible when based on well-established CF registries but issues related to quality control and missing data obviously remain crucial (van der Ent, 2008). A number of methodological issues make it even harder to compare data from different countries. These include (but are not limited to) the population coverage level, the choice of reference equations, the type of subjects included in the registry (an increasing number of patients with milder forms of the disease are now being identified in some countries through wide access to CFTR gene sequencing and nasal potential difference measurements, or via CF NBS), a lack of uniform definitions of specific items (FEV 1 has been recorded as the last, the best, the mean or the average of the best value for each quarter of the year; normal FEV 1 has been defined as  80% or  90% of the predicted value in different registries) or of clinical conditions (hepatopathy, CF-related diabetes, pulmonary exacerbation), differences in age stratification (in the UK, adults are defined as patients  16 years old), inclusion or exclusion of lung transplanted patients ('another disease'), wide heterogeneity of CFTR mutations throughout the world, etc. There is indeed an obvious need for standardization in data collection if we are to compare different registries meaningfully. The ever present need to 'always try to do better' that animates so many clinicians and paramedics involved in caring for CF patients has led to the development of several QI initiatives. The latter have tended to improve key indices concerning either nutrition or respiratory function, but can also have other objectives such as a greater involvement of patients as partners in care, by sending them for example a copy of their own medical records (Treacy et al., 2008). Guidelines and evidencebased medicine provide the directions in which changes for QI should be made. Various strategies can be put into place as part of these QI initiatives (Quinton, 2004;VandenBranden, 2004;Schechter, 2004;Quinton et al., 2007;Britton et al., 2008;Schechter et al., 2010;Kraynack et al., 2009;Quon et al., 2011). They all rely on adequate infrastructures, the acknowledgement that changes are needed, relevant and quantifiable objectives identified on the basis of the centre's known clinical outcomes, the deployment of the necessary means and an objective evaluation of the various steps undertaken. Taking into consideration identified obstacles to optimal treatment (Zemanick et al., 2010), trends of medication use over time (Konstan et al., 2010), and the considerable differences in treatment practices (Borsje et al., 2000) or in clinical approach between specialized CF centres (Kraynack et al., 2011) could nurture reflection and help to identify and better discern the objectives and means to be put into place.

Tools and challenges
The range of responsibilities of a team dedicated to CF patients is vast, in keeping with the complexity of this chronic disease. Close and comprehensive follow-up enables an early start of an efficient symptomatic treatment. Guidelines and specific databases belong to 12 those tools that are essential to the clinicians. The importance of the latter can not be overemphasized (Kerem et al., 2005;Quinton et al., 2007;Leal et al., 2007;Tiddens, 2009). a b c d Fig. 3. Specific database provide instant access to critical information at the individual level and at the level of the clinic as a whole. a. Current treatment modalities occasionally lead to sustained very long-term improvement of both FEV 1 and BMI. b. They can also allow prolonged survival in stable conditions despite severe pulmonary lesions (this patient eventually benefited from a lung transplant following recurrent life-threatening haemoptysis). c: This screen summarize bacteriological findings of a given patient over 5 years, also allowing to determine at a glance the state of Pseudomonas aeruginosa infection according to a meaningful classification (Lee et al., 2003). d: Instant access to key parameters at the scale of the whole clinic.
In the setting of the clinic, visualizing with a simple click trends of essential parameters such as FEV 1 or BMI can help the patient to fully understand the need for treatment modifications. www.intechopen.com

Challenge 1: Postpone chronic colonisation by Pseudomonas aeruginosa
The reasons for which PA has a predilection for the lungs of CF patients are still unclear (Gibson et al., 2003). Its presence can often be detected early, at times within the first few months of life, even in the absence of any symptom. It will usually be isolated within the first 3 years (Burns et al., 2001;Dakin et al., 2002;West et al., 2002;Hilliard et al., 2007;Sly et al., 2009;Stafler et al., 2011). A prospective study has indicated that 90% of CF children aged 4 years and above presented at least 1 positive culture for PA (Li et al., 2005). The early isolates of PA are generally non-mucoid and antibiotic susceptible. However, PA tends to colonise the airways of CF patients in a chronic, irreversible, fashion often associated to a change to a mucoid phenotype. At this stage, pulmonary function usually worsens and clinical symptoms become evident. Early detection of PA is crucial as there is a window of opportunity for effective eradication at this stage (Koch, 2002;Li et al., 2005).
Chronic PA colonisation is associated with a lower FEV 1 in childhood , an accelerated rate of FEV 1 decline (Pamukcu et al., 1995;Kosorok et al., 2001;Emerson et al., 2002), a shorter median life expectancy (CFF Registry, 1996;Emerson et al., 2002) and much higher treatment costs (Baumann et al., 2003). Preventing this colonisation is considered the most important challenge for the CF clinician, as it frequently determines the patient's future quality of life and long-term survival (CF Trust 2002;Koch, 2002). The current approach relies on two strategies (Frederiksen et al., 1999): i) paying attention to segregate patients on bacteriological grounds in order to limit the risk of cross-infections (West et al., 2002;Conway et al., 2008), ii) early antibiotic treatment at the time of the first PA colonisation (Littlewood et al., 1985;Valerius et al., 1991;Frederiksen et al., 1997;Lee et al., 2004). A number of regimens have been evaluated (Stuart et al., 2010) but there is no consensus about the best combination, dosage, or length of treatment course. An initial treatment protocol combining nebulised colistin with oral ciprofloxacin for 3 months is widely used. Combining this approach with intravenous antibiotics has also been reported, with a 5-year failure rate of only 12% (Douglas et al., 2009). A prerequisite for this approach is a close bacteriological follow-up, with a time lapse between visits that cannot exceed 3 months. This interval is shorter (up to monthly) in many centres. An aggressive approach based on repeated BAL has not proven useful (Wainwright et al., 2011). The 20% failure rate of the more commonly used approach (Frederiksen et al., 1997;Lee et al., 2004;Tacetti et al., 2004) underlines the necessity to develop other intervention modalities.
It is paradoxical that prophylactic antibiotic therapy directed against PA has only been the subject of 1 retrospective study (Heinzl et al., 2002) while prophylaxis against Staphylococcus aureus, whose threat is yet much easier to manage, has been investigated and hotly debated. The Austrian study was in fact very encouraging but the same group reported that longterm gentamicin inhalation in CF children was associated with reversible raised urinary Nacetyl-beta-D-glucosaminidase (NAG) activity, consistent with subtle subclinical renal tubular damage (Ring et al., 1998). Tobramycin and amikacin however have lower renal toxicity than gentamicin and long-term use of high doses of inhaled tobramycin (TOBI) are now considered safe (Prober et al., 2000). There are several theoretical arguments that make long-term inhaled prophylactic antibiotic therapy attractive (Lebecque et al., 2008). Based on the use of low doses of Tobramycin or Amikacin, this approach has progressively been put into place at our centre over the past 20 years, and has probably contributed, along with other factors, to a distinctly low rate of chronic colonisation rate by PA in patients under 18 years of age (<5% for more than 10 years, according to Lee's definition) (Lee et al., 2003).

Challenge 2: Maintain adequate nutrition
The poorer prognosis associated with being relatively underweight has long been recognized in CF children (Kraemer et al., 1978). An overall parallelism exists between respiratory function and nutritional status progress over time (Zemel et al., 2000;Steinkamp et al., 2002;Konstan et al., 2003;Milla, 2004;Pedreira et al., 2005). While severe pulmonary disease seriously compromises the maintenance of a satisfactory nutritional status, proof of the opposite has also been demonstrated: maintaining a good nutritional status promotes respiratory function preservation. Evidence of the possible influence of maintaining a good nutritional status was first suggested by a comparison of the CF centres of Boston and Toronto (Corey et al., 1988). In Toronto, CF patients had better respiratory function and were better fed, and the main difference in management appeared to be the lack of fat restriction in the Canadian hospital. Maintaining an adequate nutritional status has become a priority of CF care.

Many other challenges…
Early detection and adequate management of poor clinical course find their place here, along with appropriate management of complications, particularly those linked to an accelerated rate of FEV 1 decline as in ABPA (Kraemer et al., 2006), diabetes (Milla et al., 2000), gastro-esophageal reflux (Levy et al., 1986), colonisation by Burkolderia cenocepacia (Ledson et al., 2002;Courtney et al., 2004) non-mucoid strains in particular (Zlosnik et al., 2011), or by MRSA (Dasenbrook et al., 2008). In the same vein it is worth insisting on the vulnerable period spanning from pre-adolescence to adolescence . The rate of FEV 1 decline is faster at that time, and the various reasons for which this is considered a risk period should especially mobilise the attention and energy of the care management teams (Segal, 2008). In the context of CF, ABPA diagnosis is often delicate, (de Almeida et al., 2006;Thia et al., 2009) as many features overlap with those of infective exacerbations in CF, but it is important: failing to recognize ABPA can lead to irreversible lesions, whereas overtreatment exposes the patient to deleterious side-effects of systemic corticosteroids.

www.intechopen.com
The Prognosis of Cystic Fibrosis -A Clinician's Perspective

15
The prevention of cross-infections is a daily preoccupation that predetermines the detailed organisation of the CF centres. One of the major challenges is to limit Burkholderia cepacia complex (BCC) infections. The prognostic significance of colonisation by B. cenocepacia is particularly dreaded, as its presence is associated to an accelerated rate of FEV 1 decline and a clinical picture that is often fatal (cepacia syndrome) (Isles et al., 1984) and can affect patients up to then in very good health. In addition, B. cenocepacia is a cause for increased mortality following lung transplantation (Aris et al., 2001;Boussaud et al., 2008). Other germs of the BCC group such as B. multivorans or B. dolosa have a less gloomy effect on prognosis but have exceptionally also been associated to the occurrence of a cepacia syndrome (Zahariadis et al., 2003;Blackburn et al., 2004;Kalish et al., 2006).
Several other essential factors are involved in global care management. There is no equivalent long-term illness to CF today in terms of the heavy burden of its symptomatic treatment. In a recent US study, adult patients -of whom only half had performed airway clearance -reported a mean time spent on treatment activities of 108 minutes per day (Sawicki et al., 2009). Though very difficult to assess (Modi et al., 2006), treatment adherence is undoubtedly a key issue in CF (Eakin et al, 2011) and requires permanent assessment and support (Pendleton, 2000). There is no magic recipe, but this subject has to be bridged at every consultation, in an open manner, with empathy and as part of a 'therapeutic alliance' (Lask, 1994;Cohen-Cymberknoh et al., 2011). Another objective is to be able to propose psychological and/or social support in response to specific situations, either linked to the disease or impacting on treatment in real life.

Flags or indications for treatment intensification
Several clinical or biological markers have been associated with poorer outcome. Their main interest is that they can point to the need for reconsidering and optimizing symptomatic treatment. Some of these associations may appear simple common sense but the reality is often more complex: in-depth statistical analysis can reveal multiple independent factors and it is important to bear in mind that a statistically significant association is not equivalent to the demonstration of a causality link. For instance, a recent Canadian study showed that adult patients with CF who experienced at least 3 pulmonary exacerbations per year over a 3-year follow-up period were clearly high-risk patients, who warranted timely consideration for lung transplantation (de Boer et al., 2011). However, such patients were more likely to be female and diabetic, two risk factors linked to poorer outcome. In addition, while frequent pulmonary exacerbations may predispose a patient to lung transplantation, patients on the list for lung transplants may be more aggressively treated, thereby appearing to have more exacerbations. Raised serum IgG levels (Wallwork et al., 1974;Matthews et al., 1980;Wheeler et al., 1984;Proesmans et al., 2011) and the presence of localized auscultation anomalies  also belong to the list of 'flags' for treatment intensification.

Environmental factors 6.1 Socioeconomic status
As for all chronic diseases, CF prognosis is more sombre in patients of poor socio-economic status (SES). In 1986, there were 113 deaths registered of CF patients in the UK. Median age of death was 17 years overall, but it was above 20 for patients whose parents had a non-manual job whereas it was below 10 for the others (Britton et al., 1989). In a study of CF patients in the United States, the adjusted risk of death for indigent patients who qualified for Medicaid was 3.65 times higher than for those not receiving Medicaid. The average FEV 1 of Medicaid patients was less by 9.2% predicted than that of non-Medicaid patients, a difference which slightly increased by 0.54% per year of age (Schechter et al., 1998). In a CFF study of 23,817 white patients diagnosed before the age of 18 years, a strong association was found between the median household income and the mortality rate. At 6 years of age, the absolute differences in mean FEV 1 and weight percentiles from the lowest to the highest income category were already very significant and they persisted into adulthood (O'Connor et al., 2003). Furthermore, this study also clearly showed that the relationship between outcome and SES was incremental, rather than dichotomous and only affecting the most indigent. Though access to health insurance is much better in many European countries than in the US, the overall costs of CF treatment and follow-up are so high that patients with low SES are particularly at risk of inadequate resources and poor adherence. A better understanding of SES-related disparities and its causes is necessary to clarify the respective roles of a link between poverty and other environmental factors, health behaviours and limited access to optimal care. Low SES has recently been associated with lower healthrelated quality of life in CF patients and parents. After accounting for the effects of disease severity and SES, a negative effect of membership to a racial or ethnic minority on social and emotional functioning was also evident (Quittner et al., 2010).

Exposure to smoking
Exposure to passive or active smoking has deleterious effects in healthy subjects and in patients with respiratory diseases. An accelerated rate of FEV 1 decline with age is documented in healthy adult smokers. Functional repercussions of passive smoking were assessed in a large study concerning 812 CF patients (Collaco et al., 2008). Over a fifth of them (188, i.e. 23%) were exposed to passive smoking. At the age of 20, the mean FEV 1 of exposed patients was, independently of SES, 8% lower than that of non-exposed patients. This effect was even more pronounced in patients who were not homozygous for the F508del mutation, and was twice higher in patients who were also carriers of an unfavourable genotype with respect to the TGFß 1 modifier gene. Deleterious effects of active smoking should a priori be more severe but have not been investigated specifically in the context of CF. Exposure to smoking is a major environmental factor and screening for it should become routine practice in CF care as several pharmacological and nonpharmacological smoking cessation aids have proven their efficacy, the best results being obtained by combining modalities and tailoring therapy (Laniado-Laborin, 2010).

Other environmental factors
In a CF registry study, exposure to ozone and annual average exposure to particulate air pollution were both associated with an increased likelihood of pulmonary exacerbations. Exposure to particulate matter with an aerodynamic diameter of 2.5 µm or less was also associated with a decline in lung function (Goss et al., 2004). Though environmental factors are not considered the main pathogenic factors in ABPA, experts have suggested that it may be worth examining the patient's environment in refractory cases (Stevens et al., 2003). Altough the mechanisms are unclear, recent published evidence supports a link between climatic conditions (ambient temperature) and lung function in CF (Collaco et al., 2011). According to the authors, a hypothetical 18 year old white male with CF (Height: 175cm) with an FEV1 of 73.5% percent living in a cold climate would be expected to have an FEV1 of 66.1% had he resided in a 17 degree (Celsius) warmer climate.
Recreational use of marijuana is common in many countries. In a study of 173 adults, this drug was used by 20% of patients (Stern et al., 1987). In the general population, such use has been associated with pulmonary manifestations (bronchitis, pneumothorax, apical bullae) (Tetrault et al., 2007;Han et al., 2010;Gao et al., 2010) that could overlap clinical and radiological signs of CF, making it challenging to suspect them. Further studies seem warranted in this field.

CFTR genotype
Depending on their repercussions on the synthesis of the CF Transmembrane conductance Regulator (CFTR) protein, mutations of the CF gene are usually classified into 5 groups (Welsh et al., 1993). Mutations of classes I, II and III lead to the total or near-total absence of functional CFTR protein, whereas those of classes IV or V are associated to residual function of the CFTR protein, corresponding to a small % of normal activity. Pancreatic insufficiency is present in over 95% of patients carrying 2 class I, II or III mutations, whereas it is only rarely observed in patients carrying at least 1 class IV or V mutation (CF genotypephenotype consortium, 1993; Koch et al., 2001). Diabetes and severe hepatic involvement are 2 important CF complications that usually only occur in patients with exocrine pancreatic insufficiency. A recent study of 505 patients registered in Israel's databases did not observe a single case amongst 139 pancreatic sufficient patients (Augarten et al., 2008). The relationship between CFTR genotype and the severity of the pulmonary involvement is much looser. On the whole, the genotype of patients carrying 2 mutations of class I, II or III is still considered high risk and is associated to earlier mortality than genotypes including at least 1 mutation of class IV or V ('low risk' genotypes) (Mc Kone et al., 2006). Similarly, although their phenotype can be extremely variable, an overall more favourable prognosis has been associated to some mutations of class IV or V. Amongst the latter, there are mutations A455E (Gan et al., 1995), 3849 +10kbC->T (Highsmith et al., 1994;Duguépéroux et al., 2005), 2789+5G->A (Duguépéroux et al., 2005), D1152H (Musaffi et al., 2006;Burgel et al., 2010), R334W (Antinolo et al., 1997), 3272-26A->G (Amaral et al., 2001). The penetrance of the R117H mutation has convincingly been shown to be very low (Thauvin-Robinet et al., 2009) and is modulated by the polypyrimidine variant in the intron 8 acceptor splice site (T7 or T5) in cis with R117H. However, FEV 1 at a given age can be extremely variable in patients sharing a same CFTR genotype and this is observed in patients homozygous for the F508del mutation as well as in patients carrying a "milder" genotype (Gan et al., 1995). In practice, the link between CFTR genotype and severity of lung disease is not tight, making CFTR genotype most often of little value in predicting the prognosis at the individual level.

Modifier genes
Every clinician has in mind extreme examples of the heterogeneity of functional respiratory progress in CF patients, including amongst those that are homozygous for the F508del mutation. Though poorly compliant, chronically colonized by PA since adolescence and diabetic for 10 years, one 40 year-old patient keeps his FEV 1 above 80% of the predicted value whereas a young lady, with an optimal follow-up since birth and only occasionally Aspergillus fumigatus in sputum experiences a rapid functional decline from age 12 and requires a lung transplantation at 20. Environmental factors may play a role but admittedly such patients have to be either 'protected' for one or 'condemned' for the other, by particular allele combinations of modifier genes probably modulating the immune and inflammatory response in the lungs.
Family-based studies and especially comparisons of monozygous and dizygous twin pairs have proven fruitful in identifying and assessing the contribution of modifier genes in CF disease. Many studies in this field have yielded conflicting results and it is now realized that the repeatability of SNP-phenotype association studies with positive findings was low when less than 500 participants were included (Boyle, 2007). More recent studies are more powerful and use more sophisticated tools including whole-genome methods. Quite convincing evidence is now available that variants of at least 3 genes can be associated with lung disease severity in CF (Cutting, 2010): i) MBL2-deficient (10q) genotype: in normal subjects, mannose binding lectin (MBL) aids the phagocytosis of bacteria and deficiency in MBL seems to predispose to early infection with PA in CF ii) Increased TGFß 1 (19q) expression (Drumm et al., 2005): this gene, which also modulates the risk for asthma and chronic obstructive pulmonary disease, encodes a cytokine playing a role in the regulation of inflammation and tissue remodelling iii) Increased EDNRA (4q) expression: Endothelin is a proinflammatory peptide and smooth muscle agonist which is increased in CF airways. Deleterious effects could be related to an impact on smooth muscle tone in the airways and/or vasculature (Darrah et al., 2010). More recently, two loci causing variations in CF lung disease severity have been identified on chromosomes 11p and 20q respectively (Wright et al., 2011). Variation in TCF7L2 (10q) was reported to increase the risk of diabetes about threefold and even more in patients without previous treatment with systemic steroids (Blackman et al., 2009). Variants of other genes are suspected to increase the risk of liver cirrhosis (Bartlett et al., 2009) or meconium ileus. Further progress in the identification of modifier genes should result in an increased ability to predict severity of CF disease, and hopefully be accompanied by new perspectives for therapeutic intervention.

Conclusions
For patients, their relatives and also their carers, facing CF is often compared to running a long-distance race. It is also a team race, where no-one can let go. In terms of life expectancy, prognosis of the disease has been improving continuously since 40 years. However, it is almost always impossible to predict CF prognosis at the individual level. The quality of an early global care management is an essential prognostic determinant. Further progress in symptomatic CF treatment remains necessary, especially towards better prevention of respiratory involvement in early-diagnosed newborns and in the field of immunosuppression in lung transplantation. The discovery of a cure for the respiratory disease in CF would be a real breakthrough. A pharmacological approach tailored to the class mutations of the CFTR gene appears currently the most encouraging route in this direction.
Meanwhile, the essence of the carer's role today remains founded on these two maxims: 'Do not abandon the marathon' and 'Always try to do better'.