Cystic fibrosis

Objectives: Previous studies have shown that female patients with cystic fibrosis (CF) have a significantly poorer prognosis than male patients. Such studies investigating gender-related differences have generally combined data from several centers. The aim of this study was to determine whether with modern aggressive treatment of CF this is still true when care is standardized within a single center.

Design: Retrospective analysis of annual assessment data constructing two cross-sectional studies for the year 1993 (56 female patients, 49 male patients) and 2002 (115 female patients, 94 male patients) and two longitudinal studies, each lasting 5 years, starting in 1993 (21 female patients, 19 male patients) and 1998 (40 female patients, 41 male patients). Outcome measures included mortality, height, and weight SD scores (z scores), and percent predicted for lung function.

Results: In neither cross-sectional study were there significant differences between the sexes for median [FEV.sub.1] percent predicted (1993: female patients, 86%; male patients, 84%; 2002: female patients, 93%; male patients, 92%). Female height and weight z scores were at least as good as those of male scores. In the longitudinal studies, there were no clear trends toward declining lung function or growth, but the overall [FEV.sub.1] percent predicted appeared to be better in female patients than male patients for both cohorts. This was statistically significant for the 1998 cohort (female median [FEV.sub.1], 91.5% [range, 28 to 134%]; male median [FEV.sub.1], 84.8% [range, 32 to 145%]; p < 0.05). Female nutritional status was at least as good as male nutritional status, other than the 1998 weight z scores (-0.54 vs -0.21, respectively; p < 0.02). Since 1993, there have been 13 deaths altogether (7 female patients).

Conclusion: During childhood and adolescence, the lung function and nutrition of CF patients should be at least as good in female patients as in male patients. Individual clinic practice should be reviewed if a gender gap persists.

Key words: cystic fibrosis; gender gap; prognosis; survival

Abbreviation: CF = cystic fibrosis

**********

Cystic fibrosis (CF) is one of the most common and lethal inherited diseases. Its natural course is one of progressive deterioration in health, with respiratory morbidity secondary to chronic inflammation being the leading cause of death. There have been dramatic improvements in life expectancy during the past few decades in CF patients worldwide due to aggressive patient management and the development of new pharmacologic agents. However, female patients with this disease have been said to have a significantly poorer prognosis and lower survival rates than male patients, (1,2) particularly in the age range of 1 to 20 years. (1,3) This could be because there is an unavoidable interaction between the primary defect and gender, which may not be ameliorated by therapy, or because one or more complications are more common in female patients. In the latter ease, if the complication could be prevented or treated, then the difference would disappear. This distinction is important, because in the latter case, clinics reporting a gender gap would need to examine their practice critically. Several potential risk factors have been suggested to explain the difference in gender-related survival such as earlier acquisition of Pseudomonas aeruginosa in female patients (2) and a more rapid decline in growth and nutritional measures in girls coincidentally with a decline in lung function. (4) Many studies in the past that have investigated gender differences have used data combined from multiple centers. This approach has the advantage of large numbers being able to be recruited, but treatment differences between centers may obscure important associations and effects. (5) Over the last 10 years, the CF database at the Royal Brompton, London, has recorded the results of the annual assessments of children with CF attending the clinic. We hypothesized that with modern aggressive treatment of CF there is no longer any difference in prognostic factors between male patients and female patients when care is standardized within a single center. Growth and lung function were the main outcome measures investigated.

MATERIALS AND METHODS

Subject Population and Data Collection

Although the Children's CF Clinic database was initiated in 1983, it was not until 1993 that the clinic was sufficiently large enough to allow meaningful statistical analysis on the data from selected children. In that year, 244 children were registered on the database. A shared care program involving the referring district general hospitals existed for 70% of subjects. By 2002, the clinic had increased in size, with 330 subjects currently registered; 80% of these subjects were in a shared care program. All subjects had received a diagnosis of CF that had been confirmed with a positive sweat test (chloride level, > 60 mmol/L [from two sweat samples]) and/or having the presence of two recognizable CF disease-causing mutations in the CF transmembrane conductance regulator gene. (6) Annual review assessment data from children old enough (ie, [greater than or equal to] 5 years age) to undergo measurements of lung function, as well as growth and microbiological data, were used. Growth and lung function were the main outcome measures investigated. Lung function was measured using a pneumotachograph (Jaeger CompactLab; VIASYS Healthcare; Conshohocken, PA) in the lung function laboratory by trained technicians. [FEV.sub.1] and FVC were measured and were expressed as percentages of predicted values obtained from equations based on gender and height by Rosenthal et al. (7) Weight (measured with minimal underclothing) and height (using a Harpenden stadiometer) were measured and were expressed as SD scores (z scores) using published population means and SDs. (8) Groups were examined for differences in age, age at diagnosis, genotype, presence of pancreatic insufficiency, and the presence of P aeruginosa in airway secretions derived from sputum or deep throat swabs obtained after coughing from nonexpectorating patients. Samples were processed in accordance with standard clinical laboratory methods for the isolation and investigation of P aeruginosa. (2) The diagnosis of pancreatic insufficiency was confirmed either by 3-day fecal fat analysis (normal, <5 g/d) and/or by the measurement of stool elastase with pancreatic insufficiency, which is classified as < 15 [micro]g/g stool.

Study Design

Differences between the genders were investigated by retrospective cross-sectional and longitudinal analyses. The cross-sectional analysis (between genders and within gender) compared outcome measures of subjects undergoing annual assessments in the year 1993 with those from 2002. The gap of 9 years from 1993 to 2002 was the longest possible and was chosen to minimize any overlap of outcome data on subjects in the 1993 group into the 2002 cohort. The longitudinal analysis tracked the data on a subgroup of children from the cross-sectional study of 1993 for whom there were sufficient regular annual assessment data for 5 consecutive years (1993 to 1997) and compared these data with those extracted in a similar manner from a group of children in 1998 (ie, tracked from 1998 to 2002). Results were again analyzed for between-gender and within-gender differences. For both cross-sectional and longitudinal analyses, genotype group comparisons were between patients homozygous for the [DELTA]F508 mutation compared against all other mutations; the group sizes were too small to permit further subdivisions. Correlation between age at diagnosis and outcome measures was explored. For the cross-sectional analyses, one or more isolations of P aeruginosa in airway secretions led to the patient being classified as infected. For the longitudinal analyses, patients were classified as infected with P aeruginosa if there was an isolate in the last year of each cohort (ie, 1997 for the 1993 cohort and 2002 for the 1998 cohort).

Statistical Analysis

The power of the study and the sample size necessary for the cross-sectional study were determined by assuming that to demonstrate a statistically significant difference between girls and boys of 10% in [FEV.sub.1] percent predicted a total sample size of 120 subjects (for a 5% significance level at 80% power) would be required. This calculation was based on determining the standardized difference (the clinically important difference between the groups divided by the SD of the predicted values). (9) This calculation assumed that the data approximated to a normal distribution. Prior to the analysis of outcome measures, box-plot graphs were constructed to identify outliers using a statistical software package (SPSS, version 6.1; SPSS; Chicago, IL). Outliers were then verified on the database and from medical records, and where necessary data were amended. For nonparametric analysis, the Mann-Whitney test was used to compare the median values between groups for age, age at diagnosis, and outcome measures. For the longitudinal analysis, median values were calculated using the total number of measures for 5 years in each cohort. The prevalence of potential risk factors, genotype, pancreatic insufficiency, and the presence of P aeruginosa colonization in airway secretions were compared using the Pearson [chi square] test. Correlation between age at diagnosis and outcome measures was investigated using Spearman rank-order correlation. A p value of 0.05 was taken as the limit of statistical significance.

RESULTS

Cross-Sectional Analyses

The 1993 group consisted of 56 female patients and 49 male patients who were compared to 115 female patients and 94 male patients in the 2002 group. Thirteen children from the 1993 group have subsequently died (7 female patients [median age at death, 18.2 years; age range, 10.42 to 21.5 years] and 6 male patients [median age at death, 21.3 years; age range, 18.4 to 23.0 years]). No deaths have occurred in the 2002 group; the data from patients who died were included in the study. The cross-sectional and longitudinal data analyzed were all gathered while the patients were < 18 years of age. There was no difference between male patients and female patients for the occurrence of pancreatic insufficiency or the presence of P aeruginosa, although there was a higher prevalence in female patients in 2002 (67%) compared to female patients in 1993 (48%; p = 0.02). There was no significant difference between the sexes or between the years for the homozygous [DELTA]F508 genotype (range, 50 to 57%). Both in 1993 and 2002, female patients appeared to be younger than male patients, but the difference was not statistically significant. In 1993, 28% of both female patients and male patients were in the adolescent age range compared to 29% for female patients and 36% for male patients in 2002. Female patients received diagnoses later than male patients (0.84 vs 0.22 years in 2002, respectively; p = 0.01). No significant correlation was found between age at diagnosis and outcome measures for either gender or year group (Tables 1, 2).

In both 1993 and 2002, there was a tendency for female patients to have better lung function than male patients, although the difference was not statistically significant. Lung function appeared to improve for both female and male patients between 1993 and 2002 (Fig 1, 2). This was statistically significant for female patients with median [FEV.sub.1] values of 86.0% predicted (range, 18 to 131% predicted) in 1993 and 92.7% predicted (range, 28 to 136% predicted) in 2002 (p = 0.05).

[FIGURES 1-2 OMITTED]

Median height z score was significantly greater in 1993 for female patients compared to male patients (0.10 vs -0.44, respectively; p = 0.01) [Fig 3], Differences in height were also observed between female patients; female patients in 2002 were significantly shorter than those in 1993, with a median height z score of 0.10 (range, -2.72 to 2.85) for female patients in 1993 and -0.44 (range, -3.9 to 1.59) for female patients in 2002. No differences were observed for weight (Fig 4).

[FIGURES 3-4 OMITTED]

Longitudinal Analyses

The 1993 cohort comprised 21 female patients and 19 male patients, and the 1998 cohort comprised 40 female patients and 41 male patients (Tables 3, 4). These numbers are lower than those in the cross-sectional groups because children could only be selected for the longitudinal studies if they had completed 5 consecutive years of formal lung function testing and growth assessment. There were several reasons that children failed to undergo consistent annual testing. Some children were too unwell, some forgot, some were reluctant to undergo the tests, and some moved to other areas of the country or abroad. Nine of the 1993 longitudinal cohort were included in the 1998 cohort. No deaths occurred during either of the 5-year study periods. Five deaths have subsequently occurred among children in the 1993 longitudinal cohort, (two female patients [ages at death, 15.4 and 16.6 years]; three male patients [median age at death, 21.6 years; age range at death, 19.5 to 22.3 years]). No deaths have subsequently occurred among children in the 1998 cohort. There was a trend toward later diagnosis in female patients, but there was no significant correlation between age at diagnosis and any of the outcome measures for female and male patients in either of the cohorts. Male patients in both cohorts received diagnoses at approximately the same age, whereas female patients in the second cohort received diagnoses 9 months later than did those in the first cohort. There was no difference between male and female patients within and between the cohorts for pancreatic insufficiency and those homozygous for the [DELTA]F508 genotype. However, there was a greater prevalence of P aeruginosa infection in male patients in the 1998 cohort (83%) when compared to female patients in the same cohort (63%%; p = 0.04) and male patients in the 1993 cohort (58%; p = 0.04) [Table 4].

Female patients in the 1993 cohort were significantly younger than male patients (median age, 10.97 vs 12.96 years, respectively; p < 0.001). In 1993, there were no female patients who were teenagers compared to 2 of 19 male patients (11%). By 1997, at the end of the 5-year study, 9 of 21 girls (43%) and 16 of 19 boys (84%) were [greater than or equal to] 13 years of age. In the 1998 cohort, 4 of 40 girls (10%) and 3 of 41 boys (7%) were in their teens, and by 2002, at the end of the 5-year study, 25 of 40 girls (63%) and 23 of 41 boys were in their teens. However, there were no significant differences for median age between the sexes. Female patients were younger in 1993 compared to female patients in 1998 (median age, 10.97 vs 11.99 years, respectively; p < 0.001), and male patients were older (median age, 12.96 vs 11.17 years, respectively; p < 0.001) [Table 3].

Figures 5 and 6 demonstrate the changing trend in median height and weight z scores year by year for each cohort and sex. While growth and lung function may appear to decline particularly for the 1993 cohort when comparing the median value for the initial year with those of the final year in each cohort, there were no statistically significant differences. However, the overall 5-year median height z score for female patients in the 1993 cohort was significantly greater compared to that for male patients in the same cohort (0 vs -0.66, respectively; p < 0.001), and was similarly in the 1998 set (0.00 vs -0.56, respectively; p < 0.001). No difference in height between the cohorts was observed for male patients (Fig 5). The overall median weight z score for male patients in the 1993 cohort was significantly lower compared to that for male patients in the 1998 cohort (-0.42 vs -0.21, respectively; p = 0.01). In the 1998 cohort, female patients had a lower median weight z score compared to male patients (-0.54 vs -0.21, respectively; p = 0.02).

[FIGURES 5-6 OMITTED]

Figures 7 and 8 demonstrate the changing trend in lung function expressed as the median [FEV.sub.1] percent predicted and FVC percent predicted values year by year for each cohort and sex. While there was no significant difference between the sexes for any 1 year for the 1998 cohort as a whole, female patients had significantly better lung function when the sum of the median values across the 5 years was compared ([FEV.sub.1], 91.5% vs 84.79% predicted, respectively; p < 0.001).

[FIGURES 7-8 OMITTED]

DISCUSSION

Lung function and nutritional status are the two main significant prognostic factors in CF patients. Children who weigh more and gain weight at an uninterrupted and appropriate rate have a better [FEV.sub.1] trajectory. (10,11) Malnutrition and growth retardation, which are common complications of CF, are known to adversely affect prognosis in association with declining lung function. (12) Analyzing data from the US Cystic Fibrosis Foundation National CF Patient Registry, Zemel et al (13) found that between 1991 and 1995 the weight and height of girls deteriorated faster than for boys, with an associated decline in [FEV.sub.1].

Our study has shown that for children with CF who are in mid-childhood there was little difference between female and male patients in lung function. In the cross-sectional analyses, a significant difference in lung function between the sexes could not be demonstrated, although female lung function appeared to improve significantly from 1993 to 2002. In the longitudinal analyses, the pattern was of better lung function in female patients for both cohorts, which was statistically significant for female patients in 1998. While it might be said that the subjects in the longitudinal studies differed from those in the cross-sectional studies in being selected subgroups (see "Materials and Methods" section), nevertheless there seems no a priori reason for assuming that in our population the lack of inclusion of those groups of subjects missing their annual assessments could adversely bias the outcome measures of the remaining population in favor of female patients These findings differ from those of previous studies. (1,11,14) Female patients have been shown to have a steeper rate of decline in lung function and poorer [FEV.sub.1] and FVC values in the age range of 6 to 15 years. (1,14) In the past, a variety of reasons has been suggested for the gender difference. Animal models (15) have shown that females mount a greater inflammatory response, which is associated with a higher bacterial load and greater weight loss than in males. The apparent changes in P aeruginosa infection rates found in our study should be interpreted with caution because microbiological screening procedures altered in the mid 1990s with the introduction of routine sputum and/or cough swabs. In the longitudinal study, we found that a significantly higher proportion of male patients in the 1998 cohort were infected with P aeruginosa compared to female patients within the same cohort. The reason for this is unknown. One previous study (2) has shown female acquisition of mucoid P aeruginosa to be 1.7 years earlier than that for male patients.

In the cross-sectional analyses, female patients in 2002 were significantly shorter than female patients in 1993. Similarly, in the longitudinal analyses height z scores of female patients were significantly lower in the 1998 cohort compared to the 1993 cohort. Male patients had significantly greater weight z scores in the 1998 cohort than female patients and were significantly heavier than male patients in the 1993 cohort. It has been shown that relative weight gain between ages 3 and 6 years is associated with better lung function, whereas relative weight loss in this age group is associated with worse lung function. (16) Prepubertal male and female patients have been shown to differ in height and weight, with better results for male patients. (13) Pubertal delay is common in CF patients even if the patient is well. However, our study did not investigate pubertal factors.

It could be argued that the apparent improvement in lung function for female patients in the cross-sectional study was a statistical anomaly, given that lung function measures are expressed as percent predicted for height. Therefore, female patients with lower heights would tend to have higher predicted [FEV.sub.1] and FVC values. Comparisons of actual lung function values rather than percent-predicted lung function values within defined age bands could have been made, but, given the wide age range of the population, there were too few children within each age band to allow meaningful comparison. Another possible reason for the better lung function in female patients in the context of poorer height z scores might be a gender difference in the use of inhaled or systemic corticosteroids. In the 2002 cross-sectional cohort and in the 1998 longitudinal cohort, female patients were significantly shorter. They were also older and closer to puberty than the earlier cohorts, at a time when corticosteroids may have a more profound effect on growth. The effect of these drugs could not be looked at in detail in this study. Information on medications was not entered into the database until 1996. This allowed the examination of their effect in the longitudinal analyses by identifying treatments in the final year of each cohort (1997 for the 1993 cohort and 2002 for the 1998 cohort). In 1997, approximately the same proportions of female and male patients received inhaled corticosteroids (female patients, 17 of 21; male patients, 15 of 19). In 2002, lower amounts of inhaled steroids were used overall, but the proportions for both sexes were approximately the same (female patients, 20 of 40; male patients, 23 of 41). The number of individuals receiving systemic corticosteroids was too small too to be significant. It would seem unlikely, therefore, that corticosteroid use could explain the height z score discrepancy between the sexes.

There was a trend toward later age at diagnosis in female patients in both the cross-sectional and longitudinal analyses. This phenomenon has been seen before and is thought to be due to the mode of presentation in female patients as they are more likely to present with respiratory symptoms alone. (17) There is some evidence in the context of asthma that respiratory symptoms are less likely to be identified and managed appropriately in female patients compared with male patients. (18)

Other factors that may have influenced growth and lung function are environmental and social issues. One study (19) showed that patients from higher social classes had a longer period of survival. Patients at the Royal Brompton come from various backgrounds and regions, but social class derivation is not recorded, and we could not examine the effect in a formal manner. However, it seems unlikely that there was a systematic difference in social class between male and female patients in all four cohorts studied.

We suggest that modern aggressive treatment regimens have tended to remove any adverse female gender bias from lung function. (20) Over the years, there has been an increase in the use of therapy with inhaled, IV, and oral antibiotics. The first isolation of P aeruginosa is treated very aggressively. Patients who are discovered to have this organism on routine cough swabs or sputum analysis are prescribed a 3-month course of two antipseudomonal nebulized antibiotics (gentamicin and colistin) and a course of ciprofloxacin. If the organism is still present at the end of this time, life-long therapy with nebulized antibiotics is recommended to the family of the child. If further chest exacerbations do not respond to a course of oral antibiotics, IV antibiotics are administered. A significant proportion of children with lung function below an [FEV.sub.1]. of 50% predicted have a routine program of IV antibiotic administration every 3 months. Thus, if the explanation for an apparently worse prognosis in female patients was earlier acquisition of this organism, gender differences might be much more obvious in clinics where vigorous eradication policies are not pursued. Over the years, patients with CF at the Royal Brompton have been monitored very closely regardless of the severity of the disease, and there has been an increase in the number of outpatient appointments for every patient. Many previous studies examining gender-related survival differences have used data that were combined from many centers. This study examined outcome data from one single center, thereby ensuring a degree of uniformity, particularly in the context of the measurement of the data, albeit at a cost of smaller numbers of patients being studied. Nonetheless, our power calculation shows that our study was of an adequate size to detect significant differences. To conclude, this study has shown that there appears to have been a significant improvement in the lung function of female CF patients at the Royal Brompton over the years. We could find little difference in lung function between male and female patients. However, female patients appear not to be growing as well in terms of height. Comparing the outcome data of male and female patients may be a useful audit tool for CF clinics; if there is a gender gap, it may suggest that aspects of practice are not optimal. There is no intrinsic reason why girls with CF, at least up to age 16 years, should not have equally good lung function and nutrition as their male counterparts.

Manuscript received September 5, 2004; revision accepted March 30, 2005.

REFERENCES

(1) Rosenfeld M, Davis R, FitzSimmons S, et al. Gender gap in cystic fibrosis mortality. Am J Epidemiol 1997; 145:794-803

(2) Demko CA, Byard PJ, Davis PB. Gender differences in cystic fibrosis: Pseudomonas aeruginosa infection. J Clin Epidemiol 1995; 148:1041-1049

(3) Kulich M, Rosenfeld M, Goss CH, et al. Improved survival among young patients with cystic fibrosis. J Pediatr 2003; 142:631-636

(4) Zemel BS, Jawad AF, FitzSimmons S, et al. Longitudinal relationship among growth, nutritional status, and lung function in children with cystic fibrosis: analysis of the Cystic Fibrosis Foundation National CF Patient Registry. J Pediatr 2000; 137:374-380

(5) Johnson C, Butler SM, Konstan MW, et al. Factors influencing outcomes in cystic fibrosis. Chest 2003; 123:20-27

(6) Rosenstein BJ, Cutting GR. The diagnosis of cystic fibrosis: a consensus statement. J Pediatr 1998; 132:589-595

(7) Rosenthal M, Bain SH, Cramer D, et al. Lung function in white children aged 4 to 19 years: I. Spirometry. Thorax 1993; 48:794-802

(8) Tanner JM Whitehouse RH, Takaishi M. Standards from maturity for height, weight, height velocity and weight velocity: British children 1965; part 2. Arch Dis Child 1966; 41:613-635

(9) Gore MS, Altman DG. How large a sample in statistics in practice. Annapolis Junction, MD: BMJ Publishing Group, 1982

(10) Peterson ML, Jacobs DR, Milla CE. Longitudinal changes in growth parameters are correlated with changes in lung function in children with cystic fibrosis. Pediatrics 2003; 112:588-592

(11) Kerem E, Reisman J, Corey M, et al. Prediction of mortality in patients with cystic fibrosis. N Engl J Med 1992; 326:1187-1191

(12) Shepherd R. Achieving genetic potential for nutrition and growth in cystic fibrosis. J Pediatr 2002; 140:393-395

(13) Zemel BS, Jawad AF, FitzSimmons S, et al. Longitudinal relationship among growth, nutritional status, and lung function in children with cystic fibrosis: analysis of the Cystic Fibrosis Foundation National CF Patient Registry. J Pediatr 2000; 137:374-380

(14) Corey M, Edwards L, Levison H, et al. Longitudinal analysis of lung function decline in patients with cystic fibrosis. J Pediatr 1997; 131:809-814

(15) Guilbault C, Stotland P, Lachance C, et al. Influence of gender and interlekin-10 deficiency on the inflammatory response during lung infection with Pseudomonas aeruginosa in mice. Immunology 2002; 107:297-305

(16) Konstan MW, Butler SM, Wold MEB, et al. Growth and nutritional indexes in early life predict lung function in cystic fibrosis. J Pediatr 2003; 142:624-630

(17) Lai HC, Kosorok MR, Laxova A, et al. Delayed diagnosis of US females with cystic fibrosis. Am J Epidemiol 2002; 156:165-173

(18) Kuhni CE, Sennhauser FH. The Yentl syndrome in childhood asthma: risk factors for undertreatment in Swiss children. Pediatr Pulmonol 1995; 19:161-166

(19) Britton JR. Effects of social class, sex and region of residence on age at death from cystic fibrosis. Lancet 1989; 298:483-486

(20) The Royal Brompton and Harefield NHS Trust: Clinical guidelines for the care of children with cystic fibrosis 2003. Available at: www.rbh.nthames.nhs.uk/childrencf. Accessed January 1, 2004

Nisha Verma, BSc; Andrew Bush, MD; and Roger Buchdahl, MD

* From the Imperial College (Ms. Verma), London, UK; and Royal Brompton and Harefield National Health Service Trust (Drs. Bush and Buchdahl), London, UK.

Correspondence to: Roger Buchdahl, MD, Consultant Paediatrician, Royal Brompton and Harefield NHS Trust, Sydney St, London SW6 6NP, UK; e-mail rbuc@easynet.co.uk

COPYRIGHT 2005 American College of Chest Physicians
COPYRIGHT 2005 Gale Group