Article Text

PDF

Risk of invasive pneumococcal disease in people admitted to hospital with selected immune-mediated diseases: record linkage cohort analyses
  1. Clare J Wotton,
  2. Michael J Goldacre
  1. Unit of Health-Care Epidemiology, Department of Public Health, University of Oxford, Oxford, UK
  1. Correspondence to Professor Michael J Goldacre, Unit of Health-Care Epidemiology, Department of Public Health, University of Oxford, Old Road Campus, Oxford OX3 7LF, UK; michael.goldacre{at}dph.ox.ac.uk

Abstract

Background Invasive pneumococcal disease is a serious infection, and it is an important cause of morbidity and mortality in certain groups of ‘at-risk’ people. Those considered ‘at-risk’ in the UK include very young children, people aged 65 years and older and people with certain serious chronic diseases, asplenia or immunosuppression. There is little evidence about whether people with immune-mediated diseases are at increased risk of pneumococcal disease and therefore may benefit from pneumococcal vaccination.

Methods Retrospective cohort studies, using linked hospital data, from the longstanding Oxford Record Linkage Study (1963–2008) and from recent English national linked Hospital Episode Statistics (1999–2008); analysis of whether people with immune-mediated diseases are more likely than others to be admitted to hospital for pneumococcal disease; calculation of rate ratio for pneumococcal disease in cohorts with immune-mediated disease compared with control cohorts.

Results There were elevated rate ratios for many of the immune-mediated diseases, for example, Addison's disease in England 3.8 (95% CI 3.4 to 4.2), autoimmune haemolytic anaemia 4.9 (4.4 to 5.3), Crohn's disease 2.2 (2.1 to 2.3), diabetes mellitus 3.7 (3.4 to 4.1), multiple sclerosis 3.7 (3.5 to 3.8), myxoedema 1.60 (1.58 to 1.63), pernicious anaemia 1.74 (1.66 to 1.83), primary biliary cirrhosis 3.3 (2.9 to 3.7), polyarteritis nodosa 5.0 (4.0 to 6.0), rheumatoid arthritis 2.47 (2.41 to 2.52), scleroderma 4.2 (3.8 to 4.7), Sjogren's syndrome 3.2 (2.9 to 3.5) and systemic lupus erythematosus 5.0 (4.6 to 5.4). Findings in the Oxford and all England data sets were similar.

Conclusions People admitted to hospital with immune-mediated diseases are at higher risk than those with invasive pneumococcal disease. Vaccination should be considered in this group of patients.

  • Invasive pneumococcal disease
  • multiple sclerosis
  • systemic lupus erythematosus
  • immune-mediated disease
  • record linkage
  • public health
  • epidemiology
  • public health policy
  • health services
  • access to healthcare
View Full Text

Statistics from Altmetric.com

Introduction

Invasive pneumoccocal disease is a serious infection caused by the bacterium Streptococcus pneumoniae. It can cause pneumonia, septicaemia and meningitis. It is an important cause of morbidity and mortality, particularly in very young children, older people and immunocompromised or asplenic patients. The current NHS immunisation programme in the UK aims to protect people who are more likely than others to get pneumococcal infection or in whom infection is likely to be more serious. Current guidelines advise vaccination for infants as part of the routine childhood immunisation programme, using a 13-valent pneumococcal conjugate vaccine. A pneumococcal polysaccharide vaccine, currently 23-valent, is given to those aged 65 years and older and to those aged 2 months or older who are in the following clinical groups: asplenic or dysfunction of the spleen; chronic respiratory diseases such as chronic obstructive pulmonary disease (COPD), chronic bronchitis or emphysema; chronic heart disease; chronic kidney disease; chronic liver disease; diabetes; those who are immunosuppressed; individuals with cochlear implants and those with cerebrospinal fluid leaks.1

Recent studies have suggested that people with some immune-mediated diseases may be at an increased risk of pneumococcal infection. A recent population-based case–control study performed in the UK found that rheumatoid arthritis and multiple sclerosis (MS) were risk factors for pneumococcal disease.2 Systemic lupus erythematosus (SLE),3 psoriasis,4 inflammatory bowel disease4 and pemphigoid5 have also been suggested as risk factors for pneumococcal disease.

The aim of this study was to use the longstanding Oxford Record Linkage Study (ORLS) and a more recent English national linked Hospital Episode Statistics (HES) data set to determine the risk of invasive pneumococcal disease in patients admitted to hospital with a range of immune-mediated diseases.

Methods

Population and data

The data sets and methods used have been described in detail elsewhere.6–10 The ORLS includes statistical abstracts of records of all hospital admissions (including day cases) in NHS hospitals, and all deaths, in defined populations within the former Oxford NHS Region. The original ORLS (ORLS1) covered the years 1963 to 1998, and a second data set (ORLS2) covered 1999–2008. Because identifiers used in the Oxford NHS Region changed in 1999, the two data sets cannot be linked to each other. Fortuitously, the periods covered by ORLS1 and ORLS2 are roughly the same as the periods prior to and after the introduction of routine pneumococcal immunisation in England. We used the complete data set of English national linked HES from its inception in 1999 to 2008. The unlinked hospital data are similar to those in administrative admission, discharge or separation systems now in use in many industrialised countries. Death data in the data sets derive from computerised death registration records. The population covered by ORLS1 gradually expanded over time from an initial population of part of Oxfordshire (approximately 300 000 resident population) to all four counties of the former Oxford NHS region (resident population 2.5 million); the ORLS2 covers the same four counties. The data sets used in this study (versions m6v2 for ORLS and v13 for ORLS2 and England) have been matched and linked using software developed by staff in the Oxford record linkage team.

The matching and linkage methods used for ORLS1 have been described in detail elsewhere.7 ,8 In brief, probabilistic matching was undertaken using patients' names, addresses, dates of birth, sex and other corroborating data. The matching and linkage methods for the English national data and ORLS2 used encrypted versions of each person's NHS number (unique to the individual), HESID (a national hospital number unique to each individual), date of birth, sex and postcode.

We describe the methods used for the analysis of rheumatoid arthritis followed by invasive pneumococcal disease. The same basic methods were used in the study of every other immune-mediated disease covered in this paper. A cohort of people with a record of admission or day case care for rheumatoid arthritis was constructed by identifying the first admission, or episode of day case care, for the condition in an NHS hospital during the study period of 1963–1998 in ORLS1, 1999–2008 in ORLS2 and 1999–2008 for the whole of England. The International Classification of Disease (ICD) codes used for each immune-mediated disease can be found in tables 1 and 2. Invasive pneumococcal disease was defined using the following ICD10 codes: septicaemia due to Streptococcus pneumoniae, A40.3; pneumonia due to S pneumoniae, J13 and J18.1; or pneumococcal meningitis, G00.1. Equivalent ICD codes were used for earlier years of the study. A reference cohort was constructed by identifying the first admission for each individual with various other, mainly minor, medical and surgical conditions (listed in table 2 footnotes), as used in other similar studies by us.9–13 We followed the practice when using hospital controls of selecting a diverse range of conditions, rather than relying on a narrow range (in case the latter are themselves atypical in their risk of subsequent disease).

Table 1

Number of people in the study with each immune-mediated disease and percentage of females

Table 2

Rate ratios* and 95% CIs for pneumococcal disease in people with selected immune-mediated diseases compared with the control cohort

People with pneumococcal disease prior to, or at the same time as, the first admission for rheumatoid arthritis or the reference conditions were excluded from the analysis. We searched the database for any subsequent NHS hospital care for, or death from, pneumococcal disease in these cohorts.

Statistical methods

We calculated rates of pneumococcal disease based on person-years. We took ‘date of entry’ into each cohort as the date of first admission for rheumatoid arthritis, or reference condition, and ‘date of exit’ as the date of first record of pneumococcal disease, death or the end of the data set (31 December 1998 for ORLS1; 31 March 2008 for ORLS2 and English HES), whichever was earliest. In the ORLS cohorts, we compared the rheumatoid arthritis cohort with the reference cohort, by calculating the rates for pneumococcal disease, stratified and then standardised by age (in 5-year age groups), sex, calendar year of first recorded admission and district of residence, to ensure that the results of group comparisons were equivalent in these respects. In analysing the England data set, we used similar methods but stratified by region (rather than district) of residence and we included strata for the quintile of patients' Index of Deprivation Score (as a measure of socioeconomic status, not available in the ORLS). We used the indirect method of standardisation, and the combined rheumatoid arthritis and reference cohorts as the standard population. We applied the stratum-specific rates in the combined cohorts to the number of people in each stratum in the rheumatoid arthritis cohort, separately, and then to those in the reference cohort, to obtain the expected number of people with pneumococcal disease in the rheumatoid arthritis cohort and in the reference cohort. We calculated the rate ratio of pneumococcal disease in the rheumatoid arthritis cohort relative to that in the reference cohort. CIs for the rate ratios and χ2 statistics for their significance were calculated as described elsewhere.14

Hospital admission for people first recorded as having diabetes when aged younger than 30 years was used as a proxy for type 1 diabetes, as type of diabetes is not well recorded in routine English hospital statistics.

In the ninth revision of the ICD, used up to 2001 in our data sets, the coding of ‘pneumococcal pneumonia’ and ‘lobar pneumonia, organism unspecified’ are equivalent: both are coded as code 481. We included the ICD10 code J18.1 for ‘lobar pneumonia, unspecified’ in our analysis. Although it is likely that a good proportion of cases coded as such are attributable to pneumococcal infection, there will be some that are not. To address this, we performed a sensitivity analysis excluding the code J18.1 from the analysis for England to compare with the main results.

In order to determine whether the risk of pneumococcal disease in people with immune-mediated diseases is comparable to the risk of well-established risk factors for which immunisation is already recommended, we additionally looked at COPD and chronic renal disease as exposure conditions.

Results

The numbers of people in the study, and percentages who were female, are shown in table 1.

There were significantly high risks of pneumococcal disease after hospital admission, in all three data sets, for Addison's disease, autoimmune haemolytic anaemia, Crohn's disease, diabetes under 30 years of age, MS, myxoedema, pernicious anaemia, polyarteritis nodosa, primary biliary cirrhosis, psoriasis, rheumatoid arthritis, scleroderma, Sjogren's syndrome and SLE (table 2, see appendix 1, online only, for numbers on which the rate ratios are based).

In the data set for all England, which was much larger, there were also increased risks for pneumococcal disease in people admitted with ankylosing spondylitis, chronic active hepatitis, dermatomyositis/polymyositis, idiopathic thrombocytopaenia purpura, myasthenia gravis, pemphigus/pemphigoid, thyrotoxicosis and ulcerative colitis (table 2 and appendix 1, online only).

Particularly high levels of risk, higher than the risks associated with some of the other diseases, were found for Addison's disease, autoimmune haemolytic anaemia, MS, polyarteritis nodosa, scleroderma and SLE.

Sensitivity analysis

Numbers of cases were significantly decreased when J18.1 was excluded from the England analysis, but rate ratios remained high and significant. For example, the rate ratio for Crohn's disease was 2.15 (95% CI 1.75 to 2.63, based on 101 cases), MS was 2.19 (95% CI 1.72 to 2.77, based on 73 cases) and rheumatoid arthritis was 3.10 (95% CI 2.84 to 3.39, based on 604 cases). The complete set of results excluding J18.1 can be found in appendix 2 (online only).

Comparison with well-established risk factors for pneumococcal disease

We additionally investigated the risk of pneumococcal disease in people hospitalised with COPD and chronic renal failure. The rate ratio following COPD was 3.14 (95% CI 2.93 to 3.36, based on 1225 observed and 507 expected cases) in ORLS1 and 4.19 (95% CI 4.14 to 4.24, 64 923 observed and 28 493 expected cases) in the England data set. The rate ratio following chronic renal disease was 2.60 (95% CI 2.14 to 3.12, based on 118 observed and 46.7 expected cases) in ORLS1 and 2.87 (95% CI 2.83 to 2.92, 22 040 observed and 9583 expected cases) in England.

Discussion

We investigated the associations between hospitalisation for a range of immune-mediated diseases and subsequent hospitalisation for, or death from, pneumococcal disease. We studied these conditions in two independent data sets, ORLS1 (1963–1998) and England (1999–2008), and in two data sets in the same geographical area, ORLS1 (1963–1998) and ORLS2 (1999–2008), which cover periods that are approximately pre-immunisation and post-immunisation periods for the use of pneumococcal vaccine in high-risk individuals. We found elevated risks of pneumococcal disease in all three populations in people admitted to hospital for Addison's disease, autoimmune haemolytic anaemia, Crohn's disease, type 1 diabetes mellitus, MS, myxoedema, pernicious anaemia, polyarteritis nodosa, primary biliary cirrhosis, psoriasis, rheumatoid arthritis, scleroderma, Sjogren's syndrome and SLE. Analysis of the risk of pneumococcal disease in people with COPD and chronic renal disease, which are well-established risk factors for pneumococcal disease, showed that the risk in people with some of the immune-mediated diseases was of similar magnitude to that associated with COPD and renal failure. This gives our findings face validity. Given that pneumococcal vaccination is currently given in the UK for these conditions, it seems reasonable to assume that risk of pneumococcal disease in people with the immune-mediated diseases that have rate ratios of a similar magnitude may be clinically important and that immunisation should be considered for these diseases too.

A recent study performed by us using the ORLS and England data found that people admitted to hospital with coeliac disease were at an elevated risk of pneumococcal disease,13 and several other recent studies have suggested associations with other immune-mediated diseases, such as rheumatoid arthritis,2 ,15 MS,2 SLE,3 psoriasis,4 inflammatory bowel diseases4 and pemphigoid.5

The years covered by our two ORLS data sets roughly fall either side of the reasonably wide introduction of the pneumococcal vaccine in the UK and can be considered to approximate a pre-vaccination era (ORLS1, 1963–1998) and a post-vaccination era (ORLS2 and all England, 1999–2008). In fact, comparing ORLS1 and ORLS2, rates are not generally lower in the latter period. This is likely to be, in large part, due to a relatively limited uptake of the vaccination by people in these disease groups. A study of the uptake of pneumococcal vaccine in at risk populations, performed in England and Wales recently, found that, for example, only 27.7% of people with diabetes in 2001 had received the vaccination in the previous 5 years, rising to 34.7% in 2005.16 The authors suggest that the rise in uptake in more recent years is likely to be, in part, due to the introduction of the pneumococcal vaccine programme for people aged 65 years and older.16 Future studies are needed to monitor whether the risk of pneumococcal infection is declining in patients with these immune-mediated disease.

A recent population-based case–control study used patient data from general practices in the UK over a period of 10 years between 1996 and 2005.2 They studied the risk of pneumococcal disease in people with the established risk factors for vaccination and sought to find any additional risk groups. As well as confirming risks in those groups already offered the pneumococcal vaccine, including those with diabetes, they additionally found elevated risks in people with rheumatoid arthritis and MS, as well as other conditions such as stroke and cancer. They found an adjusted odds ratio (OR) for pneumococcal disease in people with rheumatoid arthritis of 1.84 (95% CI 1.62 to 2.10) and for those with MS of 3.63 (95% CI 2.70 to 4.88).2 These are similar to our findings, with our equivalent rate ratios in the England data being 2.47 (95% CI 2.41 to 2.52) for rheumatoid arthritis and 3.65 (95% CI 3.49 to 3.82) for MS. Although it could be argued that people admitted to hospital with an immune-mediated disease are more likely to have more severe disease than those not admitted and may therefore be at greater risk from pneumococcal infection, the agreement of the OR/rate ratio between the two studies, the general practice study and our hospital-based study, suggests that there is an elevated risk of pneumococcal infection regardless of severity of rheumatoid arthritis or MS. Another study found a rate ratio for pneumonia in people with rheumatoid arthritis to be 1.68 (95% CI 1.46 to 1.95) in people with confirmed infection.15 Both these previous studies show rate ratios that are a bit more conservative than ours, for rheumatoid arthritis, suggesting that the elevation of risk is somewhere between one and a half and two and a half times that of the general population.

Treatment of immune-related diseases, with immunosuppressive drugs, is thought to play a role in the increased susceptibility to infection. In particular, recent biological treatments, such as tumour necrosis factor-alpha (TNFα) antagonists, that alter immune response are thought to increase the risk of infection.17 ,18 Infection has been implicated as a contributory cause of death in people with rheumatoid arthritis,19 SLE20 and inflammatory bowel diseases,21 ,22 and case reports have found associations between SLE23 and ulcerative colitis24 and subsequent severe pneumococcal infection. We lack information on any treatment used for immune-mediated diseases and are unable to determine the extent to which treatment may affect pneumococcal disease occurrence in this population. Further analyses using data sets that contain treatment or prescribing information, for example, the General Practice Research Database, may help more accurately to determine the risk of infection in this population that is attributable to immunosuppressive treatment.

Strengths and weaknesses

The data sets are large and can usefully be used to study uncommon conditions, as some of the immune-mediated diseases are. The ORLS1 data provide long duration of follow-up; the English data provide a much larger and more recent population but with shorter follow-up. The fact that we were able to study a wide range of different diseases, as ‘exposures’, within a single study design, means that the risk associated with different immune-mediated diseases can be directly compared with each other.

The data sets have limitations, as discussed in greater detail elsewhere.10 Data were not recorded on patients who move out of, or are treated in hospitals outside, the areas covered by data collection (this mainly affects the ORLS studies). We have had to assume that migration rates in the immune-disease cohorts are acceptably similar to those in the reference cohorts. These factors were part of our considerations in including a comparison cohort of patients admitted to hospital, or in receipt of day case care, from the same database and for ‘matching’, through stratified analysis, for area of treatment and for year of first recorded diagnosis as well as for age and sex. The data sets are limited to people who were admitted to hospital or who received day case specialist care. We did not have access to clinical and laboratory data, and information on treatment. For all these reasons, our results should be regarded as suggestive rather than definitive. Our findings on the risk of pneumococcal disease in people with COPD and renal failure, as mentioned above, do nonetheless give face validity to the results on immune-mediated disease. Further work is needed to confirm or refute the findings on immune-mediated disease. Studies involving direct patient contact would be difficult to undertake on the required scale, but other large data sets such as the General Practice Research Database and the Health Improvement Network could be used for further investigation.

The size of the data set for England is such that many findings are significant, even if the elevation of risk is fairly small (eg, ulcerative colitis, myxoedema, thyrotoxicosis), as a result of high statistical power.

In summary, given the findings from previous studies and the results reported here, invasive pneumococcal disease appears to occur more frequently than expected in this population of patients. Extension of vaccination programmes to include a range of immune-mediated diseases should be considered.

What is already known on this subject

  • Only a few studies have examined the association between different immune-mediated diseases and pneumococcal infection.

  • Single studies and case reports have noted elevated risks of pneumococcal disease in people with rheumatoid arthritis, MS, systemic lupus erythematosus and ulcerative colitis.

  • To our knowledge, no study has addressed a wide range of immune-mediated diseases using the same methodology and data sets.

What this paper adds

  • We used the Oxford Record Linkage Study and the English national linked Hospital Episode Statistics data sets to determine if individuals with a range of immune-mediated diseases had an increased risk of developing pneumococcal disease.

  • We found a significant increase in the incidence of pneumococcal disease among individuals with many immune-mediated diseases, including Addison's disease, autoimmune haemolytic anaemia, MS, polyarteritis nodosa, rheumatoid arthritis, scleroderma and systemic lupus erythematosus, within a UK population.

  • Vaccination should be considered for this group of patients.

Acknowledgments

Over many years, the linked datafiles were built by Leicester Gill, Matt Davidson and Myfanwy Griffith, Unit of Health-Care Epidemiology, University of Oxford.

References

View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

Footnotes

  • Funding The Unit of Health-Care Epidemiology is funded by the English National Institute for Health Research to analyse the linked data. The views expressed in this paper do not necessarily reflect those of the funding body.

  • Competing interests None.

  • Ethics approval Ethical approval for analysis of the record linkage study data was obtained from the Central and South Bristol Multi-Centre Research Ethics Committee (04/Q2006/176).

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.