Incidence, socioeconomic deprivation, volume-outcome and survival in adult patients with acute lymphoblastic leukaemia in England

Acute lymphoblastic leukaemia (ALL) is a haematological malignancy arising from lymphoblasts. Whilst it is a relatively common malignancy in children with a higher incidence in boys, the incidence is rare in adults. The prognosis is good in childhood ALL but is poor in the adult form of the disease [1]. The epidemiology of childhood ALL has been well documented but much less has been done on the epidemiology in adults.

Socioeconomic inequalities in health are a major cause of concern and socioeconomic deprivation is associated with the incidence and prognosis of many conditions [2]. With regard to the incidence of ALL in childhood, a review of early studies found that higher incidence appeared to be associated with higher socioeconomic status [3]. However, another review which included several subsequent studies found conflicting evidence [4]. Little has been done on the incidence of ALL in adults in relation to socioeconomic deprivation and a study of young adults found no evidence of association [5].

The association between socioeconomic deprivation and survival has been examined in a number of studies on childhood ALL, with poorer survival associated with deprivation in several studies but not others [5‐10]. Little research, however, has been carried out on the association between socioeconomic deprivation and survival in adults with ALL, with no significant association found in young adults [5].

The association between the volume of patients treated at a hospital and outcome has significant implications for the organisation and centralisation of services and has been investigated in relation to an increasing range of conditions, including medical and surgical care for cancer patients [11]. With regard to ALL, centres treating a high volume of childhood ALL cases were found to have better outcomes in some studies but not others [12‐14]. However, the association between volume and outcome has not been investigated in relation to ALL in adults.

Trends over time and geographical variation are other aspects relevant to the epidemiology of ALL in adults. In one region in the UK, there appeared to be no improvement in survival of young adults diagnosed between 1990 and 2002 [5]. However, studies in other countries which included more recent years have reported improvements in survival over time [15, 16]. International and within country geographical variation in ALL incidence and survival has been examined, although mainly in children or with all ages combined [1, 5, 10]. Little, however, has been done on the differences in incidence and survival between urban and rural areas in relation to adults with ALL.

We examined the incidence and survival of adults with ALL in relation to age, gender, socioeconomic deprivation, rurality and trends over time. We also examined the association between volume of patients treated by hospital providers and survival.

Whilst we found no evidence of association between socioeconomic deprivation and incidence of ALL in adults, survival with higher amongst patients living in affluent areas. Hospitals treating low volumes of adults with ALL were associated with poorer survival. Rurality did not appear to be associated with incidence or survival. Incidence was higher in men but there was no evidence of a gender difference in survival. Survival improved over time but was poor amongst older patients.

A review of early studies of childhood ALL found that higher incidence was associated with higher socioeconomic status [3]. However, a subsequent review found that the evidence was mixed, with several studies reporting lower incidence of childhood ALL in more affluent groups [4]. A more recent study found a non-significant decrease in incidence with increasing affluence in analyses combining children and adults [22]. Another study found no evidence of association between incidence and deprivation in young adults, consistent with our results for all adults [5].

With regard to survival, a national study in the UK on childhood ALL found that higher levels of socioeconomic deprivation were associated with poorer survival [9]. The authors argued that the association was not due to different treatment across social groups as the association was also seen in children enrolled in the UKALL XI trial. Our results indicate that socioeconomic inequalities in survival are also a significant cause for concern in adults with ALL and further research is required to understand and address this issue.

Our finding that mortality was higher for adults with ALL treated at hospitals treating low volumes of adult ALL patients is of concern. An early study on volume and outcome in the treatment of childhood ALL found that centres treating an average of six or more patients per year had the best long-term survival rates [12]. Although two recent childhood ALL studies found no association between volume and outcome, they only examined mortality during induction therapy [13, 14]. A study of in-hospital mortality in patients undergoing chemotherapy for acute myeloid leukaemia, a disease predominantly of adults, was the first to report that the mortality rate was significantly higher in low-volume hospitals [23]. High volume centres in general are more likely to have specialised clinical teams and more experienced staff to manage patients with specific medical conditions, which could explain better outcomes [24].

Our finding of improved survival over the time period examined is consistent with studies of adults with ALL in Germany, the USA and the Netherlands [15, 16]. We found no gender differences in survival. Others have reported that whilst survival was worse in men previously, the gender difference appears to no longer be apparent [15]. Survival, however, remains poor, especially amongst older adults and those with relapsed or refractory disease [1, 25].

Our study has a number of potential limitations which need to be considered. We used HES data which has the advantage of being a comprehensive national system and all adults with ALL are likely to have been admitted at least once to confirm the diagnosis. However, coding and data entry errors could have led to over or underascertainment of cases. We examined the data in detail in order to arrive at a set of procedures to keep errors to a minimum. The average case counts we obtained were comparable to the ONS estimates from cancer registration data. In addition, the five-year survival of 36% we observed for patients receiving chemotherapy is consistent with the five-year survival of 38% for adults in the UKALL 12 trial [26]. There may have been errors in the mid-year population estimates we used to calculate incidence rates. We used an area based measure of socioeconomic deprivation assigned to individual patients and some patients may have been misclassified. However, we used a measure based on small geographical areas which would have minimised this error as levels of socioeconomic deprivation for populations in smaller geographical units are likely to be generally more homogenous. The volume indicator was based on a relatively small number of cases because of the limited time span for which we had consistent provider unit codes, and some providers may have been misclassified by volume. However, misclassification in this context might have diluted the strength of the association. Potential errors need to be taken into account in the interpretation of our results.

In terms of future research and policy, further work is needed to confirm the association between higher socioeconomic deprivation and poorer survival and the observation that hospitals treating a low volume of patients were associated with poorer outcome for adults with ALL. These findings, if confirmed, are likely to have significant implications for the organisation of services for the treatment of adults with ALL.