Discussion
This study found that neonatal complications, low FEV1/FVC, low FEF25-75 and low body mass index were associated with increased mortality in Indian children with cystic fibrosis. European cystic fibrosis society patient registry database of 2007 suggested that low BMI, chronic pancreatic insufficiency, chronic colonisation with Pseudomonas and development of cystic fibrosis-related diabetes mellitus increases the risk of having poor lung function as demonstrated by poor FEV1 value (8). However, our study patients did not have a statistically significant incidence of CFRD, chronic colonisation with Pseudomonas or chronic pancreatic insufficiency. Moreover, as per the natural history of the disease, the complications like CFRD usually develop in the second decade of life. In our study population, most children had died just after the first decade of life at an average age of 13.5 years (95% CI- 8.3 -17.0).
Earlier studies done in adults reported that low FEV1 was associated with increased mortality (8). It has been demonstrated by Kerem et al. that patients with an FEV1 value < 30% predicted to have a 40- 50% chance of dying with two years in a cohort followed up for 12 years (9). FEV1 was the strongest predictor of mortality in a study conducted by Huang et al., although the authors mentioned that other lung function test indices were not strong predictors of mortality (10). There was a trend of low FEV1 in the died group in our study (Table 1), but it was not a predictor of mortality in COX regression analysis. A possible explanation may be early death (average age at death 13.5 years) in our cohort before they had very low FEV1. FEV1 is effort dependent and FEV 25-75 is relatively less dependent on patient’s effort. However, low FEF25-75 and the low ratio of FEV1/FVC is a relatively novel finding of the present study.
The case fatality rate in our study was 18.0%. A study done by Courteny et al. indicated an average case fatality rate of 24.6 in adults (11). Lower BMI and lower PFT results are in general associated with higher mortality reported from various studies from different parts of the world (12-14), as seen in our study.
The average age at death was 13.5 years in our study, which is much lower than the average life expectancy of CF patients in western countries (15). This is possibly due to the early diagnosis and availability of CF specific drugs in western countries. CF in developed countries is no longer a paediatric disease than developing countries where it still qualifies to be a disease of children. Most children had died in the first half of the second decade of their lives. This shortened life span in developing countries like India is attributed to delay in diagnosis and lack of proper medical facilities even after diagnosing disease. Because of the lack of national guidelines and support, most children are treated symptomatically without giving them advanced care in CF specific drugs. Since the cost of CF specific drug therapy is exorbitantly high, most parents cannot bear the cost of treatment and hence prefer symptomatic medical/surgical management as necessary. Average age at which symptoms appeared first and at which diagnosis of CF was confirmed was 6 months (95 %- 1.5,24) and 60 months (95%- 17,120) in the group that had died during the period of the study. Even in our study, the time lag between the onset of symptoms to diagnosis of CF was around two to four years. This delay is attributed to lack of suspicion at primary/secondary care centres due to lack of awareness, non-availability of diagnostic tests, and further delay in referral to higher centres equipped with CF specific care. Most children had poor nutrition status at the diagnosis itself. Family history of CF was found in very few children in our study, and it did not contribute to increased mortality.
We had more representation of male children in our study cohort; however, the death rate at the end of the study period was higher in female children (Figure 1). This corresponds to the earlier studies conducted by Kerem et al. (9) and Rosenfeld et al. (16) in the adult population. Although the exact cause as to why female children have higher mortality is not known to us, it may be related to better nutritional status and better medical care given to male children in India. However, further research is required to know the exact cause of higher mortality in female children because even in developed countries where no preferential treatment is given to male children, mortality is reported higher in female children (9,16).
Homozygous Delta F 508 is the commonest mutation in western countries and is associated with higher mortality. A mutation study revealed that heterozygous delta F508 is the most common mutation in this study cohort. An earlier study done by Connor et al. found that heterozygous delta F508 mutation was associated with a reduced risk of mortality compared to homozygous delta F508 mutation (17). Another study done by Johansoen et al. (18) suggested that homozygous delta F508 has a higher risk of mortality than other mutations. If we compare the studies of Connor et al. and Johansoen et al., it is evident that the presence of heterozygous delta F508 mutation was associated with lower mortality. However, there was no association between the type of mutation and the risk of mortality in our study. This is possible because we could not do a complete mutation study of the entire study cohort due to the lack of facilities and heterogeneous mutation profile in our population.
BMI was a significant predictor of mortality in our study. This is again corresponding to an earlier study by Kerem et al. (9), Bell SC et al. (19) and Fired et al. (20). However, Courteny et al. conducted a study on predicting mortality in adult CF patients and found no correlation of increased mortality with BMI (11).
There was no significant association of any lab parameter with the mortality on bivariate analysis. Although airway colonisation was seen in almost 2/3rd of our study population, it did not contribute to overall mortality. The most common bacteriological agent for colonisation was Pseudomonas, followed by staphylococcus and Burkholderia. Age of colonisation and the organism of colonisation had no bearing on overall mortality in our study. This contrasts to earlier studies where early colonisation and colonisation with species like Burkholderia were found to significantly contribute to mortality in patients with CF (20-24). A possible explanation for not finding an association between airway colonization and increased mortality in our study may be, that other factors like poor nutrition and not been able to afford the treatment contributed more to early mortality.
CF-related complications like ABPA, CF-related liver disease and Pulmonary artery hypertension was found in very few children, and it did not contribute to mortality.
The strengths of our study are that it was conducted over a large cohort that was followed over ten years and is likely to represent the disease-specific traits in the Indian population. This is one of the first few studies with a relatively large number of patients where we have tried to analyse the predictors of mortality in Indian children with CF.
There are few limitations also of our study. Few data were missing due to the retrospective nature of the study. We had no access to newer CF specific drugs like lumacaftor, Ivacaftor and Tezacaftor due to lack of funds, which might have some survival benefits. Finally, we could not do a complete genetic analysis of all the patients.