Objective To review and synthesise evidence on rates of respiratory-associated deaths and associated risk factors in the intellectual disability population.
Design Systematic review and meta-analysis.
Data sources Embase, CINAHL, ISI Web of Science (all databases including Medline) and PsychINFO were searched for studies published between 1st January 1985 and 27th April 2020 and examined study and outcome quality. Reference lists and Google Scholar were also hand searched.
Results We identified 2295 studies, 17 were included in the narrative synthesis and 10 studies (11 cohorts) in the meta-analysis. Data from 90 302 people with intellectual disabilities and 13 808 deaths from all causes in people with intellectual disabilities were extracted. Significantly higher rates of respiratory-associated deaths were found among people with intellectual disabilities (standardised mortality ratio(SMR): 10.86 (95% CI: 5.32 to 22.18, p<0.001) compared with those in the general population, lesser rates for adults with ID (SMR: 6.53 (95% CI: 4.29 to 9.96, p<0.001); and relatively high rates from pneumonia 26.65 (95% CI: 5.63 to 126.24, p<0.001). The overall statistical heterogeneity was I2=99.0%.
Conclusion Premature deaths due to respiratory disorders are potentially avoidable with improved public health initiatives and equitable access to quality healthcare. Further research should focus on developing prognostic guidance and validated tools for clinical practice to mitigate risks of respiratory-associated deaths.
PROSPERO registration number CRD42020180479.
- respiratory medicine (see thoracic medicine)
- public health
- respiratory infections
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Strengths and limitations of this study
To the best of our knowledge, this is the first systematic review and meta-analysis on respiratory-associated deaths among people with intellectual disabilities.
Included studies were limited by sample size.
There was no sufficient data and results provided by studies to investigate predictors or factors associated with respiratory-related deaths; meta-regression or stratification was not possible.
The meta-analysis included mortality ratios from 10 observational studies covering 90 302 people with intellectual disabilities and 13 808 deaths from all causes in people with intellectual disabilities.
A rigorous and systematic analysis process was undertaken which minimised the risk of bias, errors and omissions.
People with intellectual disabilities account for approximately 1%–3% of the global population.1 2 The World Health Organisation (WHO)3 defines intellectual disabilities as impairments in adaptive functioning, social functioning and intellectual functioning (IQ<70), requiring a need for daily support, with the onset in the developmental phase (<18 years). While some heterogeneity is to be expected in the definition of intellectual disabilities across studies drawing on administrative data sets, the WHO definition can be applied to all studies included in this review. Life expectancy and mortality rates are important indicators of health inequality.4 People with intellectual disabilities die up to 20 years earlier than the general population.5–8 Respiratory disorders are a leading cause of death among people with intellectual disabilities.6 9 The range of standardised mortality ratios (SMRs) due to respiratory disorders for people with intellectual disabilities are very high in some studies,10–12 and much lower in others.13–15 Despite this, SMRs due to respiratory disorders for people with intellectual disabilities differ widely across studies. Respiratory cause of mortality in people with intellectual disabilities has not been systematically examined. Previous studies have focused on either children and young people (4–19 years)10 or older adults (55+years) on average.12 This systematic review and meta-analysis aims to investigate and quantify the risk of, and factors associated with, respiratory-associated deaths in people with intellectual disabilities.
The Preferred Reporting Items for Systematic Reviews and Meta‐Analyses checklist was followed.16 This review was prospectively registered with the International Prospective Register of Systematic Reviews.
This systematic review included studies which analysed and presented data on people who were ascertained as having intellectual disabilities and a comparison group of individuals in the general population, with respiratory disorders included as a separate cause of death. For studies that included multiple disabilities, at least 70% of participants had to have intellectual disabilities, if results were not reported separately. Studies also had to be full-text, peer-reviewed and published in English. To be included in the meta-analysis, studies had to report SMRs with 95% CIs for respiratory-associated deaths based on external comparison group or to have presented data allowing such outcomes to be derived. Studies were excluded if they focused on specific etiologies of intellectual disabilities, such as Down syndrome, as these are associated with different health and mortality profiles compared with other people with intellectual disabilities. Studies were excluded if the full paper was not available in English. Studies focussing on postoperative and post-treatment deaths were excluded as these are not representative of the wider population with intellectual disabilities. Studies with small samples (<20 participants) or case series designs were also excluded as these papers are less representative.
Search strategy and selection criteria
We searched Ovid Embase, ISI Web of Science (all databases), CINAHL and PsycINFO from 1 January 1985 to the 27 April 2020, using comprehensive terms related to ‘intellectual disabilities’, ‘mortality’ and ‘respiratory disease’ (full search strategy in online supplemental appendix 1). In addition, a manual bibliography and citation search of included studies was conducted using Google Scholar and key researchers in the field of mortality in individuals with intellectual disabilities were emailed to identify any additional relevant papers. The aforementioned eligibility criteria were used. After duplicates were removed, all records were imported into Covidence software (www.covidence.org) for title and abstract and full-text screening. All titles, abstracts (CM and AMcG) and full-texts (CM, AMcG and ER) were double-screened with inter-rater reliability (Cohen’s kappa) of ĸ = 0.57 and ĸ = 0.58, respectively.
Data extraction was conducted using a structured database created in Excel. Five researchers (GSS, LAH-M, DK, KD and AMcG) each extracted data from 25% of the included studies and, to check reliability, one other researcher (CM) independently extracted data from 20% of included studies. Extracted data were compared in meetings and discrepancies resolved through consensus discussion. Researchers did not extract data on included papers where they were a listed author.
Assessment of study and outcome quality
Study quality was appraised using the Standard Quality Assessment Criteria for Evaluating Primary Research Papers from a Variety of Fields.17 Quality ratings were calculated in percentage form using the standard method17 and categorised as weak (<55%), moderate (55%–75%) or strong (>75%) quality. Each paper had quality appraisals completed by two researchers, who then agreed a consensus score for each item (table 1).17 Researchers did not evaluate quality of papers where they were a listed author. Risk of bias score was not used to exclude any studies from either the systematic review or meta-analysis. We evaluated the quality of our own systematic reviews using the Measurement Tool to Assess Systematic Reviews checklist.18
Summary of outcomes and statistical analysis
Findings of all included studies were combined in a narrative synthesis. The primary goal of the meta-analysis was to investigate if the SMRs of respiratory-associated deaths differ for individuals with and without intellectual disabilities. If SMRs were reported by specific respiratory causes, sex, age group, level of intellectual disability, socioeconomic status or ethnicity, these were collected and presented for potential analysis (see table 2). Random-effects meta-analysis was undertaken using RevMan. Included studies reported either:
an SMR or HR
The observed number of deaths or expected deaths necessary to calculate a SMR. These were calculated using STATA V.14 by dividing the observed number of deaths in a cohort study group by the expected mortality based on age and gender-specific death rates in the general population comparison group.
Random-effects models were selected for all meta-analyses due to the different populations and measures in the included studies. Inverse of the variance method was used to calculate the weighted mean respiratory mortality log-SMR across studies, as well as for subgroup meta-analyses. As the SMR is a ratio, log transformation was needed to maintain symmetry in the analysis.19 SMRs and HRs from each study were transformed to log values for computations and back transformed for presentation of the results. Weighted mean log-SMRs and their 95% CIs were reported separately for individuals with and without intellectual disabilities. The magnitude of the back transformed ratio and associated CI were also reported. Where data permitted, further subgroup analyses were conducted to examine sources of heterogeneity. Where more than two studies reported subgroup level data, or cause-specific results of causes of respiratory deaths (eg, pneumonia) random-effects models were considered for subgroup meta-analyses.
For the random-effects meta-analysis, heterogeneity was expected in the pooled result. Therefore, the χ2 statistic I2 was chosen to measure level of heterogeneity across the studies, as it allows for interpretation of results regardless of the number of studies included in the meta-analysis, the type of outcome data, or effect measurement.20 Heterogeneity was interpreted as not observed when I2=0%, low when I2=25%, medium when I2=50% and high when I2=75%.20
Sensitivity analysis was used to assess the impact of risk of bias for each study on the weighted mean SMR. Data were removed one-by-one from the meta-analysis for each study, beginning with the lowest ranked papers, to determine their effect and re-estimate the weighted mean SMR. Cumulative analysis, starting with larger studies and sequentially adding smaller studies, was used to investigate how the weighted mean SMR estimate changes as small studies are added.21
Patient and public involvement
No patient and public involved.
Figure 1 summarises the systematic search, selection and reasons for exclusion. All 17 studies were included in the narrative synthesis and 10 were included in the meta-analysis (studies with relevant SMRs n=8 and HR n=2). A full list of studies excluded from full-text screening is available in online supplemental appendix 2.
Table 1 illustrates the characteristics of studies reporting mortality rates for respiratory disorders and pneumonia in people with intellectual disabilities and table 2 presents all-cause mortality and deaths from respiratory disorders in people with intellectual disabilities.
Key features of all studies identified for inclusion in the review were tabulated (table 1). These were cohort studies (n=12), case–control studies (n=4) and one population-based audit of deaths in adults and children. These studies report data on 90 302 people with intellectual disabilities and 27 394 deaths. The average study size was 9250 people. These studies were from the Netherlands (n=1), Finland (n=2), Australia (n=3), the UK (n=7), the USA (n=1), Sweden (n=2) and Denmark (n=1).
Definition of respiratory disorder
Thirteen out of 17 (76%) studies defined the respiratory disorder using International Classification of Diseases (ICD) 9-chapter codes13–15 22 23 and ICD 10-chapter codes for respiratory disorders.8 10 12 23–26 The remaining four studies included in the systematic review did not define respiratory disorders.
Causes of death from respiratory disorders
Thirteen papers reported on cause of deaths from respiratory disorders.8 10–15 22–24 26–28 Pneumonia was reported as a cause of death in 12 studies,8 10–15 22–24 26 27 five studies reported deaths from pneumonitis related to aspiration,8 10 12 14 24 five studies reported on chronic obstructive pulmonary disease (COPD),11 12 14 26 27 one study reported on asthma29 and one reported respiratory cancer deaths.8
Five papers reported that respiratory disorders were the dominant cause of death in people with intellectual disabilities.11 13 27 28 30 A further three studies found that deaths from respiratory disorders were the second most common cause of death.12 14 24 Respiratory-associated deaths were in the top five main causes of deaths for a further four papers.9 10 22 25 Comparative results (intellectual disabilities vs general population) for deaths due to respiratory disorders were reported in 10/17 (59%) of the studies.10 12–15 22 24 27 28 30 In the majority of these studies, rates of death from respiratory disorders were higher for people with intellectual disabilities than for people in the general population. However, Troller et al25 reported that respiratory-associated deaths in the general population were (9%) similar to the population with intellectual disabilities (12%). Hollins et al11 also reported that respiratory disorders were the most commonly cited cause of death for both groups.
Individual respiratory disorders and mortality
Pneumonia was reported as the most common cause of respiratory death in people with intellectual disabilities.8 10–15 22–24 26 27 Contributors to pneumonia deaths included influenza and injury from inhalation and aspiration events.10 14 Pneumonitis featured as an underlying or contributing cause for between 8% and 21% of respiratory-associated deaths in people with intellectual disabilities.8 10 12 Crude comparison data showed people with intellectual disabilities were much more likely (between 10 and 20 times) to die from pneumonitis.24 26 COPD was found to be a common cause of death in two studies focussing on older adults.12 27
Factors associated with respiratory-associated deaths experienced by people with intellectual disabilities
Age, gender and severity of intellectual disability have been found to be associated with risk of respiratory cause of death. Only four out of 17 (23.5%) papers directly reported on factors associated with the risk of respiratory-associated deaths14 22 23 30 (see table 2). Two reported SMRs separately for males and females,14 23 while two reported proportions of respiratory deaths between males and females. None directly compared males versus females or reported tests of significance. While one study reported higher respiratory SMRs among females,23 another study reported separate SMRs for different age-bands which varied widely.14 Group-level analysis was not possible. Level of intellectual disabilities was only reported as associated with respiratory related deaths in one study with 35-year follow-up using relative risk but failed to report confidence or p-values.14 This study found that, when compared with the general population, the relative risk of respiratory related deaths was 2.6 times higher for people with mild intellectual disabilities and 5.8 times higher for people with profound and multiple intellectual disabilities.
Respiratory mortality among children and young people
Respiratory deaths among children and young people with intellectual disabilities were reported in five studies and found to be a common cause of death across all studies.10 13 15 30 31 Four studies included comparison with the general population for respiratory causes of death, while one included the national population without intellectual disabilities.10 All analyses were limited by the small numbers of death. Raitasuo et al reported only one death.13 Patja et al reported higher SMR for males aged 2–19 years but not females.31 Smith et al reported 8% deaths had respiratory disease as the underlying cause but the SMR for underlying cause was not reported.10
Ten studies8 10–15 23 24 28 reported the necessary data to calculate (SMR, HR or data necessary to calculate these) and were included in the meta-analysis of respiratory mortality of people with intellectual disabilities and the general population. As Hollins et al reported the SMR of two separate cohorts, these are displayed separately in the relevant forest plots.11 The pooled SMRs for respiratory mortality between people with intellectual disabilities and the general population was 10.86 (95% CI: 5.32 to 22.18). The results indicate that respiratory mortality occurs almost 11 times more frequently in the intellectual disabilities group than in the general population group. At the individual study level, this was adjusted for age in all studies and for sex in all studies except for two of these studies,11 13 where this was not clear. There was evidence of considerable statistical heterogeneity between studies in the meta-analyses, with I2=99.0%. Results are displayed in figure 2.
As five studies12 15 23 24 28 focused on adults only, one study10 focused on children only and six8 11–15 included people of all ages, a subanalysis was conducted of studies which reported data on an adult only population. The results of this subanalysis are displayed in figure 3. The pooled SMR reduced slightly from 10.86 (95% CI: 5.32 to 22.18) to 6.53 (95% CI: 4.29 to 9.96), after one study with a sample of primarily children was excluded.10 Studies which included both adults and children in their sample8 11–15 were next removed one at a time. First, both cohorts from Hollins et al were removed and the pooled SMR was reduced by around half, from 915 to 4.80.11 The further removal of studies by Glover et al,8 Patja et al14 and Raitasuo et al13 resulted in a final pooled SMR for adults of 5.85 (95% CI: 4.73 to 7.22, p<0.001). Heterogeneity between studies was also reduced from I2=99% to I2=56% by the exclusion of samples which included children.
A subanalysis was conducted of studies which reported an SMR for pneumonia.8 11 23 The pooled SMR for pneumonia mortality for people with intellectual disabilities compared with the general population was 26.65 (95% CI: 5.63 to 126.24, p<0.001). These results, displayed in figure 4, indicate that pneumonia-related mortality occurs much more frequently in people with intellectual disabilities than in the general population group. Evidence of considerable statistical heterogeneity between studies was also present in this subanalysis with I2=99.0%. SMRs were recalculated excluding the only study to include an adult only sample, Tyrer and McGrother23 resulting in a substantial increase in pooled SMR (95% CI: 26.65 to 42.70).
Sensitivity analysis in relation to quality assessment was run for the 10 studies included in the meta-analysis (online supplemental appendix 3). Studies which were rated as weak13 or moderate23 were removed from the analysis. The pooled SMR for mortality ratios changed slightly as Raitasuo et al13 (from 10.81 to 12.67)27 and then Tyrer and McGrother (from 12.67 to 13.94)23 were removed from the analysis. As the change in SMR was small, this suggests that inclusion of weaker studies did not significantly change the results.
This systematic review and meta-analysis highlights that people with intellectual disabilities experience excess respiratory-associated deaths, with a respiratory mortality of almost 11 times greater than for the general population. Respiratory mortality was more prevalent among studies which include children, and pneumonia was a major contributor to the higher respiratory mortality reported in this study. Clinical guidelines have contributed to a reduction in mortality from community-acquired pneumonia.29 We believe the evidence presented here highlights the need for clinical guideline development groups to make recommendations on reducing the risks of premature death due to community-acquired pneumonia among people with intellectual disabilities. Vaccination programmes for influenza can help to reduce respiratory mortality in children32 and adults.33 Although there is a relatively low uptake of influenza vaccine among people with intellectual disabilities, annual health-checks for people with intellectual disabilities have been reported to increase uptake of influenza immunisation.34 People with intellectual disabilities should be identified as a high-risk group and immunisation providers should prioritise the improvement of vaccine uptake, for example through the roll-out of health checks. People with intellectual disabilities are at increased risk of recurrent chest infections which are secondary to dysphagia35 36 with a high proportion of aspiration pneumonia-related deaths occurring among individuals with severe and profound intellectual disabilities.5 22 30 35 37 Increased recognition of the link between dysphagia and respiratory disorders among caregivers and practitioners is critical to ensuring the early identification of individuals with respiratory disorders.
The higher risk of death from respiratory disorders, such as pneumonia, for people with intellectual disabilities is a significant concern in relation to the rapidly developing COVID-19 pandemic.38 39 Urgent action to disaggregate data on deaths from COVID-19 for people with intellectual disabilities and to investigate factors associated with COVID-19-related mortality for people with intellectual disabilities is vital to ensure that clinical guidelines are based on consideration of the specific risks faced by people with intellectual disabilities. Research is urgently required to investigate the risk factors associated with COVID-19 for people with intellectual disabilities to ensure carers and clinicians have access to the best evidence to reduce the risk of infection in those most vulnerable and to inform the clinical management of those who contract COVID-19. Carers and clinical staff must be given training to ensure they understand the human rights and healthcare needs of people with intellectual disabilities to ensure that existing stark disparities in the health of people with intellectual disabilities are not widened during this crisis.
Interventions should focus on the paediatric age group. Among the studies included in this meta-analysis, we found a relationship between inclusion of children and SMRs from respiratory causes, with those studies including children reporting higher SMRs. This is consistent with studies that have reported higher SMRs in children compared with adults in epilepsy15 and cerebral palsy.40 Overall, mortality in childhood is very low relative to adulthood, and in the paediatric age group, chronic disabling conditions such as intellectual disability, epilepsy and cerebral palsy all have a marked impact on SMR. Comorbidity with epilepsy and cerebral palsy are likely to be significant modifiers of the relationship between intellectual disability and respiratory mortality. Children with more severe intellectual disability are more likely to have epilepsy and cerebral palsy, both of which are independent risk factors for respiratory mortality.
Study strengths and limitations
Our study has several strengths. The meta-analysis included mortality ratios from 10 observational studies covering 1844 respiratory deaths in people with intellectual disabilities, which has improved the power and precision to answer this important research question. A rigorous and systematic analysis process was undertaken, and we minimised the risk of bias, errors and omissions by having two or more reviewers conduct comprehensive searches, assess study quality and extract descriptive data. Due to the low prevalence (~1%) of intellectual disabilities among the general population, low sample size was a considerable limitation, relative to other patient groups. However, our meta-analysis included two national,10 12 and five regional intellectual populations in their respective countries.11 15 23 28 While heterogeneity was found, due to methodological and clinical diversity including study design, age and study nationality, this is common in meta-analyses and statistical heterogeneity was inevitable.20 We have not included assessment of non-reporting or publication bias. Most of the research was conducted in Western countries, thus limiting the extent to which the findings may generalise to non-Western countries. Furthermore, ethnicity was not reported widely which prevented further analysis. There was variation among studies on how mortality was examined and how deaths were reported. There is a general lack of evidence on factors associated with the increased risk of respiratory-related deaths in people with intellectual disabilities. As a consequence, we were not able to perform meta-regression on predictors or factors reported in studies which increase SMRs for respiratory deaths (age, sex, place of death or severity of intellectual disabilities). This should be a priority for future research in order to inform the development of targeted interventions to prevent respiratory-related deaths. Although the meta-analysis enables synthesis of data from a large sample, many of the individual studies reported on small samples and are at increased risk of bias. It is encouraging that there have been several larger studies in recent years and future research should focus on reporting respiratory mortality in representative, population-based samples. Furthermore, the majority of the studies included for review relied on death certificate data. One the most reported causes on the death certificate of people with intellectual disabilities is the intellectual disability itself. Given that this problem only exists within this population, true causes of death remain underestimated.41 42 As reporting has improved over the years, and many counties implemented automated coding systems, it is likely that older paper have more bias than more recent studies.
These findings signify the urgent need to develop and implement evidence-informed strategies to reduce premature mortality among people with intellectual disabilities. Respiratory disorders are a major cause of death for people with intellectual disabilities, many of which are avoidable with improved public health initiatives and access to good quality health and social care. However, further research is required to understand both the multifactorial causes of this heightened risk as well as the most effective approaches for the multiprofessional clinical management of these risks.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
Twitter @ScotLDO, @lah021_laura, @bhauteshjani
Contributors MT, CM, AMcG, ER, LAH-M, DK, KD, GSS, AH and FB had full access to all the data, contributed to the systematic review and meta-analysis of studies, interpretation of results and the manuscript. JS and BJ helped to interpret the result of the study and contributed to the manuscript. MT is study guarantor. All authors reviewed the final manuscript and agreed to be accountable for all aspects of the work and approved the final manuscript for submission. The corresponding author attests that all listed authors meet the authorship criteria and that no others meeting the criteria have been omitted.
Funding This work was supported by The Scottish Government grant number (HSCD 169698).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
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.