STATISTICS

Dr Sarah Walters has written a very clear article dealing with survival statistics which can be recommended to non-statisticians ("Survival in cystic fibrosis - what do the statistics mean?") the full text of which is on her website (www.cfstudy.com)  and also on the CF Trust website (www.cftrust.org.uk).

1966 Mantle DJ, Norman AP. Life tables for cystic fibrosis. BMJ 1966; 2:1238-1241. [PubMed]
First of a number of publications on survival in the UK from Dr. Archie Norman’s unit at Great Ormond Street, London. Life tables for 1943-1964 showed that 80% died by 5 years and 90% by 10 years. Infants with meconium ileus, 25% died by 1 week and 75% by 3 months.
Sydney Gellis, in an editorial comment in the corresponding Year Book of Pediatrics, observed that “Despite claims to the contrary, cystic fibrosis of the pancreas continues to carry a gloomy prognosis. Present day therapy is helpful but offers relatively little, and a realistic alteration of the course of the disease will require a major breakthrough in discovering the aetiology”. Gellis, although a distinguished paediaitrican was not an expert on CF amd in this prediction was obviously quite wrong for the outlook improved steadily over the years long before the aetiology was identifed during the Eighties.

1967 Pugh RJ, Pickup JD. Cystic fibrosis in the Leeds region: incidence and life expectancy. Arch Dis Child 1967; 42:544-545. [PubMed]
This paper is included as it is one of the few reports from the UK of a survey between 1952 and 1962 from the North of England by Yorkshire paediatricians Richard Pugh of Hull and Douglas Pickup of Pontefract on behalf of the Leeds Regional Paediatric Club. During this decade 132 infants were confirmed as having CF among total of 546,764 births over 13 years – an incidence of 1/4142. Fifty six infants (42%) died in the first year.
The findings would be typical of the situation in the UK at the time where virtually all children with CF were treated at local hospitals and there were no specialised CF clinics in most towns and cities – or even in most provincial teaching hospitals. So few patients survived childhood. A significant number of affected infants must have gone unrecognised in this study for the incidence in this part of the UK is around 1/2400 births.

1967 Young WF, Jackson AD. The prognosis of cystic fibrosis. Bibliotheca Paediatrica 1967;86:350-352.
Winifred Young and Tony Jackson were paediatricians at the Queen Elizabeth Hospital Hackney where there was one of the few CF clinics in the UK which was started by Winifred Young in 1950. Margaret Mearns joined the staff there soon after.
Further data from these authors is described above in Cystic Fibrosis. A symposium. Report of a meeting on 28th May 1964 at the Wellcome Foundation London. Chest and Heart Association.

1967 Warwick WJ, Manson S. Life table studies of mortality. In Cystic Fibrosis Part I. Physiology and pathophysiology of serum secretion clinical investigation and therapy. Proc 4th Int Conf on CF. Geigy Symposium Berne/Grundewald 1966 (Modern Problems in Pediatrics Vol 10) PFI p353 Ed. Rossi E, Stoll E. Karger: Basle and New York. [Conference report]
In 2001 Warren Warwick commented “After 10 years of grants funding the US CF Patient Registry and indirectly supporting the Canadian CF Patient Registry and the beginning of the International Registry, the CF Foundation decided the Registry should become the property of the CF Foundation. When the CF Foundation took over the operation they discovered what a bargain the $10,000 annual grant was. So after spending over $100,000 and still not finishing the analysis, they decided to give up the analysis as too expensive. They were shamed into doing the analysis when the Canadian Foundation provided funds and manpower to complete the analysis. The fallout was that the Canadian Foundation went its own way and beginning the International Registry was abandoned. All the data from the first 10 years of the US Registry, including the magnetic tapes, were transferred to the CF Foundation

1968 Warwick W, Pogue RE. Computer studies in cystic fibrosis. In Lawson D (ed). Proceedings of 5th international Cystic Fibrosis Conference, Cambridge 1969; 320-330. London. Cystic Fibrosis Research Trust. [Conference]
Warren Warwick has pioneered databases in CF and was responsible for starting the CF Foundation Patient Registry. He maintains that US paediatricians initially agreed to cooperate with his first CF database as they did not believe Leroy Matthews excellent results and thought the data may confirm their suspecion! (also Warwick & Manson, 1967).

1971 George L, Norman AP. Life tables for cystic fibrosis. Arch Dis Child 1971; 46:139-143. [PubMed]
Archie Norman and colleagues produced a number of life tables for people with CF in the UK. Here 128 infants without meconium ileus (MI) and 43 with MI from 1964 for five years are compared with tables for previous 20 years showing obvious improvement in survival – suggested as being due to early diagnosis and treatment. The one year survival was 34% compared with 93%, and at five years 19% vs. 89% between the two periods.

1973 Taussig LM, Kattwinkel J, Friedewald WT, di Sant'Agnese PA. A new prognostic score and clinical evaluation sytem for cystic fibrosis. J Pedaitr 1973; 82:380-390. [PubMed]
Often referred to as the NIH Score, this system by Lynn Taussig (figure 3.1) and colleagues stood up to evaluation well and apparently Shwachman and Kulczycki realised that a "more accurate method of assessment might become available" than their score. Respiratory function tests were included also complications that affected prognosis such as haemoptysis, cor pulmonale and pneumothorax. Some considered that the nutritional state was under represented. This and a number of subsequent scoring systems have been discussed by a number of authors (Cystic Fibrosis Clinical Scoring Systems. Conway SP, Littlewood JM. In Cystic Fibrosis - Current Topics. Volume 3. Dodge JA. Brock DJH, Widdicombe JH, editors Chichester: John Wiley & Sons Ltd, 1996:339-358.

Dr Lynn Taussig worked with Paul di Sant'Agnese at the NIH, then as Professor at the Arizona College of Medicine for 20 years and then at the National Jewish Research Center in Denver. He was particularly active in CF care and research during the Seventies and Eighties. In 1984 he edited one of the major textbooks on the CF (Cystic Fibrosis. Lynn Taussig. New York Theime-Stratton Inc. 1984). In the introduction to the book Charles Lobeck describes Lynn Taussig, Thomas Boat and Robert Schwartz as the professional descendants of Paul di Sant'Agnese who had in turn influenced their colleagues from Cleveland, Tuscon, Rochester and other academic locations. The book contains notable chapters on the gastrointestinal system, the pancreas and the hepatobiliary sytem by Paul di Sant'Agnese. There many other distinguished contributors who have made significant contributions to our knowledge of CF.

1976 Stern R, Boat TF, Doershuk CF, Tucker AS, Primiano FP Jr, Matthews LW. Course of cystic fibrosis in 95 patients. J Pediatr 1976; 89:406-411. [PubMed]
From the LeRoy Matthews and Carl Doershuk team at Cleveland reporting the long term results of their comprehensive system of management (Matthews et al. J Pediatr 1964; 65:558-575 above): these were the 95 patients they reported in 1964 after a mean follow up period of 14 years (minimum 13 years). Of the 45 diagnosed before extensive lung damage (Figure 18 Group1) only one had died and none were disabled. Of the other 50 diagnosed after substantial lung damage (Figure 18 Group 2) 26 had died – mortality being greater in females. Factors contributing to better prognosis were considered to be early diagnosis, aggressive management and comprehensive care, easy access to specialised care and improved antibiotic therapy.
The observations of this group (figure 19) on achieving successful treatment include - early inpatient treatment before significant lung damage in optimal facilities, as and when needed with flexible arrangements for daily absences, outpatient appointments throughout the week including Saturdays. Follow-up exclusively by a centre physician – usually the same physician each visit. All decisions are made by a CF centre physician.

1979 Phelan PD, Allan JL, Landau LI, Barnes GL. Improved survival of patients with cystic fibrosis. Med J Australia 1979; 1:261-263. [PubMed]
Three hundred and twenty patients with CF born after 1958 who were treated at Royal Children’s Hospital, Melbourne had an 80% survival to 11 yrs and a 64% survival to 18 yrs. Those managed between 1973-77, 91% survival to 12 yrs and 80% to 17 yrs. These were really excellent results for the time and were attributed, almost certainly correctly, to all patients having received Specialist CF Centre care as was later discussed in a very influential subsequent paper from Melbourne the contents of which prompted the formation of the UK Working Party on Cystic Fibrosis in 1982 (Phelan and Hey, 1984 below).

1984 Phelan P, Hey E. Cystic fibrosis mortality in England & Wales and in Victoria Australia. Arch Dis Child 1984; 59:71-83. [PubMed]
The contents of this paper had a major influence on CF treatment in the UK. In Victoria, Australia one death in 125 between the ages of 1 and 14 years was in a child with CF – in England and Wales the frequency was 1 in 44. A newborn child with CF had an 80% chance of surviving to 9 years in England and Wales and to 20 years in Victoria (figure 27). The better survival in Victoria during the Seventies was attributed to the Australian patients being treated at CF Centres rather than at by general paediatricians at their local hospitals. During the Seventies most UK children with CF were treated at their local hospital by general paediatricians all of whom saw only a few children with cystic fibrosis.
The authors note “It has been claimed in North America that management in a specialist centre, as in Victoria, is essential for optimal care and recent data from Denmark indicate a much better survival among patients attending a specialist clinic compared with those looked after by their local paediatrician, as happens for most children in England and Wales”.

The data in this report was available before publication through John Dodge who had visited Australia before its publication and the availability of the data was directly responsible for the formation of the British Paediatric Association’s Working Party on Cystic Fibrosis in 1982 whose terms of reference were “to assess the advantages and disadvantages of regional centres for cystic fibrosis”. For summary of the Working Party Report see Jackson 1986 (below).
John Dodge comments that he considers the original data in this report were seriously flawed because they compared mean age at death (UK) with standard survival calculations (Australia), but the point they erroneously made was of great value in setting up not only CF centres in the UK but also a CF registry, without which any data are totally unreliable!

Dr Edmund Hey, the co-author, is a UK consultant paediatrician with previous experience in neonatal physiology and paediatrics who, in 1973, became the first full time consultant respiratory paediatrician at Great Ormond Street Hospital, London; there he met Peter Phelan (figure 26) who spent a six month sabbatical there. In 1978, on the death of Gerald Neligan (the neonatal paediatrician in Newcastle, UK), Hey was invited to return to Newcastle, as neonatologist which he did. Later in 1982, Phelan invited Hey back to Melbourne to cover the sabbatical leave of a colleague; it was during this six months that Hey was involved in the present paper. When he returned to Newcastle, Hey concentrated on developing neonatal services there but later said to me (2008) “but I did watch the slow transformation of CF care across the UK with some knowledge that I had probably helped to nudge it into an important change of direction.” This is undoubtedly true!

1988 Wood RE, Piazza F. Survival in cystic fibrosis: correlation with treatment in three cystic fibrosis centres. 10th International Cystic Fibrosis Congress, Sydney 1988 Excerpta Medica Asia Pacific Services 74:79.
An impressive and important, if disturbing, study presented as a poster at the 1988 Sydney CF meeting. The medical records of three US CF Centres for the years 1971-1980 were reviewed including 24,000 outpatient visits and 2000 hospitalisations. Median survival ages at the three centres were 9.5, 18.1 and 22.8 years for centres A, B and C respectively. No differences in the content of treatment between centres correlated with group survival. But for each patient year in A, B and C the mean number of outpatient visits was 4.4, 8.0 and 12.1 and the number of hospital days was 4.7, 5.0 and 9.5. – usually hospitalisations would be for IV antibiotics. Hospitalisation was considered necessary in 19.3, 20.9 and 31.2% of patients at least once a year. There was strong correlation between group survival and the use of oral and aerosolized antibiotics, vitamins E and K, and aerosols of epinephrine. When the data was separated into 1971-1975 and 1976-1980, changes in antibiotic use correlated well with improved survival. The authors concluded “more frequent outpatient visits, more aggressive use of antibiotics and more frequent hospitalisation correlate well with increasing survival”.
It is disappointing that this excellent study was never published. I suspect the marked differences between the CF centres were unacceptable and too sensitive data for publication at that time. Interestingly similar studies, some very recent, repeatedly show patients are better when seen more frequently and receive more antibiotics – which is exactly what one would expect!

1988 Dodge JA, Goodall J, Geddes D, Littlewood JM, Mearns MB, Owen JR, Russell G. Cystic fibrosis in the United Kingdom 1977-85: an improving picture. British Paediatric Association Working Party on Cystic Fibrosis. BMJ 1988; 297:1599-1602. [PubMed]
A further publication from the Working Party on Cystic Fibrosis chaired by Prof. John Dodge. The national survey included all patients with CF known to paediatricians, chest physicians and others between 1977 and 1985 or whose deaths were reported through death certification authorities in the United Kingdom. The number indicated an incidence of 1 in 2500 live births. Mortality was 7.6% in the first year and greater for females than males under 20 years of age. Survival improved through the study particularly in the first five years with 100 more births than deaths each year. In particular, death from meconium ileus was increasingly less frequent. It was estimated that there were some 5000 people with CF in 1985. (1982 Formation of UK CF Working Party above; Jackson 1986 for summary of the initial Working Party report, above; Dodge et al, 2007 final data, below)

Figure 27: Survival curve from the paper. With permission from the BMJ Publishing Group.

1991 Littlewood AE, Bowler IM, Littlewood JM. A method of data collection and computerisation in use at a regional CF unit. 17th European Cystic Fibrosis Conference Denmark. June 1991. Poster 94.
Much of the clinical research data and the smooth running of the Leeds CF clinic were dependent on the Administrator Mrs Christine Silburn, and the clerks and secretaries.
The computer data management had been in operation since the early Eighties using first a computer funded by the Variety Club of Great Britain. Ann Littlewood (figure 25) a state registered nurse (and my wife!), was responsible for the running of our data service from the early Eighties until we both retired from clinical work in 1997.
The introduction of a Comprehensive CF Assessment service in Leeds from 1981 generated such vast amounts of data on every patient that computerisation became essential and was introduced in 1982. The system, initially using an Epson PC AX with 80Mb hard disk and 640K using Symantec Q&A software, provided a database and word processing facilities. By 1991 four databases were in use – a patient register, clinic visit database, an inpatient database and an assessment database for the results of Comprehensive Assessments. The system had been repeatedly modified and adapted over the next eight years by interested colleagues and professional computer programmers. The Leeds CF units are now totally computerised for both clinical and research purposes - a system developed by Dr Daniel Peckham, Respiratory Physician on the Adult CF Unit.

1996 Frederiksen B, Lanng S, Koch C, Hoiby N. Improved survival in the Danish center-treated cystic fibrosis patients: results of aggressive treatment. Pediatr Pulmonol 1996; 21:153-158. [PubMed]
Survival data for Danish center-treated CF patients, covering the period 1974-1993. The annual mortality rate for 1989-1993 was 0-1.2%. Using the age-specific mortality rate for 1989-1993, it was impossible to calculate the median survival probability because the curve did not fall below 50% (age up to 45 years); however, it was possible to show that the survival probability for a newborn CF child to reach his 45th birthday was 80.4% (confidence interval 76.5-84.6%). The median age at diagnosis was 0.63 years with no sex difference (surprisingly there was and is no neonatal CF screening in Denmark). The probability of surviving 40 years after the diagnosis of CF was made was 83.3% (confidence interval 80.1-86.6%).
These survival figures are considerably better than any other published survival probability at the time. The only item that I would not agree with the Danish CF centre is their policy of not screening neonates for CF.

2000 Fogarty A, Hubbard R, Britton J. International comparison of median age at death from cystic fibrosis. Chest 2000; 117:1656-1660. [PubMed]
Data from ten countries in North America, Europe, and Australasia of all persons registered as having died of CF from 1980 to 1994 were analysed by Prof. John Britton and his colleagues from Nottingham, England. The overall median age at death increased from 8 years in 1974 to 21 years in 1994. Median age at death also increased within all countries and was consistently highest in the United States. Women were significantly more likely to die at a younger age than men.

The median age of death is one of the basic indicators of survival. It is encouraging that the median age of death steadily increases each year as the effect of better treatment is reflected in better survival. The relatively recent figures for median age of death for patients registered with the UK CF Patient Registry are encouraging - in 2002 the median age of death was 23 yrs, in 2003 24.2 yrs and in 2004 25.6 years and in 2008 it had risen to 27 years.

Dr Sarah Walters has written a very clear article dealing with survival statistics which can be recommended to non-statisticians (Survival in cystic fibrosis - what do the statistics mean?) the full text of which is on her website (www.cfstudy.com)  and also on the CF Trust website (www.cftrust.org.uk).

2000 Cheng K. Smyth RL. Motley J. O'Hea U. Ashby D. Randomized controlled trials in cystic fibrosis (1966-1997) categorized by time, design, and intervention. Pediatr Pulmonol 2000; 29:1-7. [PubMed]
Ros Smyth, Deborah Ashby and colleagues developed a register of RCTs in CF and have studied when they were performed, their design, and what interventions were investigated. They identified 506 RCTs; 37.5% were identified solely as abstract reports in conference proceedings. There has been about a 30-fold increase in the number of RCTs in CF since 1966. A high proportion of the RCTs (72.7%) had a sample size of 30 or less, and only 8.7% were multicenter trials. Reporting of study design was poor: in 51.4% the report did not state whether there was any blinding in the trial design; 53.6% of studies were of crossover design. The most common interventions studied were antibiotic treatments and physiotherapy, but a number of commonly used therapies had been evaluated only in a small number of patients. Although the number of RCTs of interventions in CF patients has increased over the last 25 years, the sample sizes of these trials are generally too small to indicate whether the intervention was effective, and very few were multicenter. Future RCTs in CF are more likely to provide clinically useful answers if higher numbers of patients are recruited into large, well-designed multicenter trials. This should be a priority of the organization of future research in CF.

2003 Johnson C, Butler SM, Konstan MW, Morgan W, Wohl ME. Factors influencing outcomes in cystic fibrosis: a center based analysis. Chest 2003; 123:20-27. [PubMed]
The analysis was conducted using data from the Epidemiologic Study of Cystic Fibrosis from 1995 through 1996. Within-site rankings tended to be consistent across the three age groups. Patients who were treated at higher ranking sites had more frequent monitoring of their clinical status, measurements of lung function, and cultures for respiratory pathogens. These patients also received more interventions, particularly courses of IV antibiotics for pulmonary exacerbations. There were substantial differences in lung health across different CF care sites. Frequent monitoring and increased use of appropriate medications in the management of CF were associated with improved outcomes.

2007 Dodge JA, Lewis PA, Stanton M, Wilsher J. Cystic fibrosis mortality and survival in the UK: 1947-2003. Eur Respir J 2007; 29:522-526. [PubMed]
Data up to 1995 on the survival of 3-yr cohorts of patients with CF born in the UK in the period 1968-1992 had previously been published (Dodge et al. Arch Dis Child 1997; 77:493-496. 9496181). The present study, coordinated by John Dodge and the last from the original Survey of the UK CF Working party, reports survival data up to the end of 2003 together with a 2003 population estimate. All subjects with CF born in the UK in the period 1968-1992 were identified up to 1997 and information from the death certification authorities up to the end of 2003 was added.
The observed survival up to 2003 of CF patients born in 1978 was 55% for males and 49% for females. For 1988 and 1992 the data were 91 and 88%, and 97 and 96%, respectively. The estimated 2003 mid-year CF population was 8,284. The continuing improvement in survival of cystic fibrosis patients in successive cohorts means that the previous prediction of median survival of >50 yrs of age for individuals born in 2000 continues to look realistic, even in the absence of proven effective therapy aimed at correcting the basic cystic fibrosis defect.

2007 Padman R, McColley SA, Miller DP, Konstan MW, Morgan WJ, Schechter MS, Ren CL, Wagener JS. Investigators and Coordinators of the Epidemiologic Study of Cystic Fibrosis. Infant care patterns at epidemiologic study of cystic fibrosis sites that achieve superior childhood lung function. Pediatrics 2007; 119:e531-7. [PubMed]
Previous analyses of the Epidemiologic Study of Cystic Fibrosis database revealed that sites with the highest average patient lung function monitor patients more frequently and treat with antibiotics more aggressively than those where average lung function is lowest. Infants at the upper quartile sites had more office and sick visits; more respiratory tract cultures; and more frequent use of intravenous antibiotics, oral corticosteroids, mast cell stabilizers (surprisingly sodium cromoglycate still seems to be used in the USA), and mucolytics; but they received less chest physiotherapy, inhaled bronchodilators, oral nutritional supplements, and pancreatic enzymes. Data shows that both enrolment characteristics and infant care patterns are associated with lung function outcomes in later childhood. The authors suggested that pulmonary function of older children may be improved through specific interventions during the first 3 years of life.

The conclusions of this analysis are likely to be obvious to those clinicians who have seen many children permanently damaged before being eventually diagnosed as having CF. Also they will have noticed the contrast of those infants diagnosed and treated early after newborn screening. However, this excellent data is very welcome to convince those who are responsible for funding CF care and neonatal screening programmes and who cannot draw on clinical experience. It is interesting that in all studies of this nature over the years the more frequent use of IV antibiotics and more frequent follow-up visits are always features in the history of those with better results (Woods & Piazza, 1988 above).