Global, regional, and national burden of neurological disorders during 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015

1In 2006, WHO emphasised the importance of neurological disorders (a group that at the time included epilepsy, Alzheimer's disease and other dementias, Parkinson's disease, multiple sclerosis, migraine, stroke, poliomyelitis, tetanus, meningitis, and Japanese encephalitis) for public health, and estimated that these disorders accounted for 6·3% of the global disability-adjusted life-years (DALYs). However, the WHO report was largely based on the Global Burden of Disease Study 2000 (GBD 2000) and did not reflect understanding of the comparative global epidemiology of neurological disorders. Increasing life expectancy and population growth worldwide in recent years mean that more people are now reaching ages where neurological disorders are most prevalent. Additionally, there have been changes in the treatment of some neurological disorders (such as the introduction of acute stroke units, thrombolysis, and thrombectomy), as well as changes in the risk factors that affect the burden of neurological disorders in the world.

2Evidence before this studyThe Global Burden of Disease Study 2015 (GBD 2015) produced comprehensive and comparable estimates of the burden of various disorders (prevalence, deaths, and disability-adjusted life years [DALYs]) for 195 countries and territories from 1990 to 2015. However, aggregated estimates in the category of neurological disorders did not include data for stroke, brain and nervous system cancer, or communicable neurological disorders, which are all of great relevance to neurologists.Added value of this studyFor GBD 2015, the GBD 2013 search strategy was replicated to take into account all relevant epidemiological studies published between 2013 and 2015. In this systematic analysis, we produced estimates of the burden of neurological disorders (stroke, Alzheimer's disease and other dementias, Parkinson's disease, epilepsy, multiple sclerosis, migraine and tension-type headache, medication overuse headache, meningitis, tetanus, encephalitis, brain and nervous system cancers, motor neuron disease, and a residual category of other neurological disorders), separately and combined. For the first time, we included stroke, brain and nervous system cancer, and communicable neurological disorders. We quantified the global disease burden (as measured by prevalence, mortality, DALYs, years of life lost, and years lived with disability) and explored variation in the burden by neurological disorder type, age, sex, and overall country development level as measured by the Socio-demographic Index (SDI). The study showed that neurological disorders ranked as the leading cause of DALYs in 2015 and the second-leading cause of death. The most prevalent neurological disorders were headaches (tension-type headache, medication overuse headache, and migraine) and Alzheimer's disease and other dementias.Implications of all the available evidenceOver the past 25 years, the burden of neurological disorders has increased substantially, and our expanded list of neurological disorders makes this the leading cause group of disability, and the second-leading cause group of mortality worldwide. Stroke remains the largest contributor to this burden globally. For all neurological disorders combined, there was a noticeable (by almost five times) gradual decrease in age-standardised DALY rates with increasing SDI, but these rates had large geographical variations. These findings could help to guide health systems and research activities to reduce the burden of neurological disorders.

3Evidence before this study

4The Global Burden of Disease Study 2015 (GBD 2015) produced comprehensive and comparable estimates of the burden of various disorders (prevalence, deaths, and disability-adjusted life years [DALYs]) for 195 countries and territories from 1990 to 2015. However, aggregated estimates in the category of neurological disorders did not include data for stroke, brain and nervous system cancer, or communicable neurological disorders, which are all of great relevance to neurologists.

5Added value of this study

6For GBD 2015, the GBD 2013 search strategy was replicated to take into account all relevant epidemiological studies published between 2013 and 2015. In this systematic analysis, we produced estimates of the burden of neurological disorders (stroke, Alzheimer's disease and other dementias, Parkinson's disease, epilepsy, multiple sclerosis, migraine and tension-type headache, medication overuse headache, meningitis, tetanus, encephalitis, brain and nervous system cancers, motor neuron disease, and a residual category of other neurological disorders), separately and combined. For the first time, we included stroke, brain and nervous system cancer, and communicable neurological disorders. We quantified the global disease burden (as measured by prevalence, mortality, DALYs, years of life lost, and years lived with disability) and explored variation in the burden by neurological disorder type, age, sex, and overall country development level as measured by the Socio-demographic Index (SDI). The study showed that neurological disorders ranked as the leading cause of DALYs in 2015 and the second-leading cause of death. The most prevalent neurological disorders were headaches (tension-type headache, medication overuse headache, and migraine) and Alzheimer's disease and other dementias.

7Implications of all the available evidence

8Over the past 25 years, the burden of neurological disorders has increased substantially, and our expanded list of neurological disorders makes this the leading cause group of disability, and the second-leading cause group of mortality worldwide. Stroke remains the largest contributor to this burden globally. For all neurological disorders combined, there was a noticeable (by almost five times) gradual decrease in age-standardised DALY rates with increasing SDI, but these rates had large geographical variations. These findings could help to guide health systems and research activities to reduce the burden of neurological disorders.

9To improve health-care planning and health outcomes of people with neurological disorders, we must understand not only the number and distribution of people with these disorders between countries, but also how these disorders affect population health (in terms of both mortality and disability) compared with other diseases and injuries.

10The Global Burden of Disease Study 2015 (GBD 2015) GBD 2015, produced comprehensive and comparable estimates of the burden (prevalence, deaths, and DALYs) of 315 diseases and injuries for 195 countries and territories from 1990 to 2015. However, aggregated estimates in the category of neurological disorders did not include stroke, brain and nervous system cancer, or communicable neurological disorders, which are of great relevance to neurologists. Ischaemic and haemorrhagic stroke were classified as cardiovascular diseases; brain and nervous system cancer as malignant neoplasms; and meningitis, encephalitis, and tetanus as communicable diseases. In this systematic analysis, we quantify the disease burden (as measured by prevalence, mortality, DALYs, and years lived with disability [YLDs]) due to neurological disorders and its relationship with development level as measured by the Socio-demographic Index (SDI; an indicator based on lagged-distributed income per person, average educational attainment among individuals older than 15 years, and total fertility rate).

11In an expanded category of neurological disorders we included stroke, Alzheimer's disease and other dementias, Parkinson's disease, epilepsy, multiple sclerosis, migraine, tension-type headache, medication overuse headache, meningitis, tetanus, encephalitis, brain and nervous system cancer, motor neuron disease, and a residual category of other neurological disorders that included diseases such as muscular dystrophy and Huntington's disease (appendix p 112).

12GBD 2015 non-fatal burden estimates were based on a systematic review of the literature to obtain all available epidemiological data on prevalence, incidence, risk of mortality, and severity. In the appendix we provide for each neurological disorder analysed the following: the list of International Classification of Diseases (ICD) codes used for mapping neurological causes of death (pp 118–21); a list of GBD sequelae, health states, health state lay descriptions, and disability weights for neurological disorders (pp 122–39); the total number of site-years by neurological cause and source type (p140); and the data representativeness index for each neurological disorder, the percentage of GBD 2015 geographies with any data by cause pertaining to the period before 2005, 2005–15, and all years of data (p 141). Reference case definitions were based on ICD-9 or ICD-10 criteria with the addition of Diagnostic and Statistical Manual of Mental Disorders (DSM)-III and DSM-IV criteria for dementia and the International Classification of Headache Disorders criteria for headaches., Sources of information used to estimate the burden of neurological disorders are on the Global Health Data Exchange website.

13Detailed GBD methods for calculating non-fatal estimates have been reported elsewhere. The epidemiological data were analysed with DisMod-MR 2.1, a Bayesian meta-regression tool that adjusts datapoints for variations in study methods between data sources and enforces consistency between data for different parameters, such as incidence and prevalence. For each neurological disorder, we defined a parsimonious set of sequelae that best described different aspects of the disabling consequences. Each non-fatal sequela was estimated separately. We calculated the YLDs caused by the residual category of other neurological disorders indirectly using a ratio of YLDs to years of life lost (YLLs). We calculated the ratio of YLDs to YLLs for Alzheimer's disease and other dementias, Parkinson's disease, multiple sclerosis, and motor neuron disease, and multiplied this ratio by the YLL estimates for other neurological disorders. Further details of the non-fatal estimates of each of the included neurological disorders are in the appendix (pp 5–113).

14Disability weights for a set of 235 health states covering all sequelae of disease and injury estimated in GBD 2015 were estimated by pair-wise comparison methods presenting pairs of lay health state descriptions to respondents in surveys done in the general population in nine countries, and an open web-based survey. The sequelae of the neurological disorders included in this analysis each map to a unique health state with a corresponding disability weight (appendix pp 122–39). YLDs were computed by multiplying the prevalence of each sequela by a disability weight and aggregating estimates for all sequelae for a disease.

15We categorised countries by their overall development status level as determined by the SDI, classifying them in high, high–medium, medium, medium–low, and low SDI quintiles on the basis of values across the 1980–2015 period (appendix pp 142–46), as described in detail elsewhere. The average expected relationships between DALY rates and death rates from neurological disorders (individually and as a group) and SDI over the entire study period (1990–2015) across all geographies in males and females were estimated using spline regression. We also categorised countries into 21 GBD regions (appendix p 155).

16GBD 2015 uses a database of 14 236 site-years of vital registration, verbal autopsy (a method of determining cause of death in countries that have no functional vital registration system), and maternal and child death surveillance data, covering the period from 1980 to 2015. Trained interviewers administer a structured questionnaire to relatives of a deceased person about their symptoms preceding death. The underlying cause of death is inferred by computer algorithms or physician review of the autopsy interview. We estimated deaths for all neurological disorders apart from headaches, to which no deaths are assigned as the underlying cause. For each neurological cause except dementia, we used the GBD Cause of Death Ensemble model (CODEm) strategy., CODEm applies mixed effects or spatiotemporal Gaussian process regression models to mortality rates or cause fractions in varying combinations with predictive covariates. The ensemble of models with best credentials on out-of-sample predictive validity tests was selected for each cause of death. YLLs were calculated by multiplying the number of deaths at each age by the standard life expectancy at that age. Results from CODEm for each disease were scaled to fit all-cause mortality estimates derived from demographic sources by location, age, year, and sex.

17We decided to use a natural history model for dementia because of a large inconsistency between the data for prevalence and mortality. For instance, in the USA, the rates of death from dementia increased five times between 1990 and 2014, whereas the available prevalence and incidence data showed no significant changes over the same period. Large increases in death rates assigned to dementia have also occurred in some other countries with high-quality vital registration systems. Furthermore, in GBD 2015, the prevalence of dementia varied among 187 countries by a factor of three, whereas dementia death rates varied by more than 20 times. The likely explanation was a change in coding practices between countries and within countries over time. To correct for this source of measurement bias, we assessed the most recent data from 23 high-income countries with high-quality vital registration systems and the highest ratio of registered dementia death rates to prevalent cases. This ratio is equivalent to the excess rate of mortality in cases of dementia. We derived a pooled estimate by age and sex using a linear regression of the log of these rates. We added these estimates as data in DisMod-MR 2.1 to derive estimates of cause-specific mortality rates that were consistent with prevalence data and the pooled estimate of excess mortality from the 23 countries that in their most recent year of vital registration were most willing to code a death to dementia as the underlying cause.

18DALYs were computed as the sum of YLLs and YLDs for each country, age, sex, and year with 95% uncertainty intervals (UIs) based on the 25th and 975th values of the ordered 1000 draws. Unless explicitly mentioned otherwise, all rates were age-standardised using the GBD standard. This study is compliant with the Guidelines for Accurate and Transparent Health Estimates Reporting (GATHER; appendix pp 114–17).,

19The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding authors had full access to all the data in the study and had final responsibility for the decision to submit for publication.

20In 2015, the neurological disorders included in this analysis caused 250·692 (95% UI 229·080 to 274·654) million DALYs, comprising 10·2% of global DALYs, and 9·399 (9·095 to 9·714) million deaths, comprising 16·8% of global deaths. We were unable to report the combined prevalence. Classifying stroke under neurological disorders rather than cardiovascular diseases made neurological disorders the largest cause group of DALYs and the second-largest behind cardiovascular diseases in terms of global deaths. Among the neurological disorders, stroke, migraine, meningitis, Alzheimer's disease and other dementias, and epilepsy each caused more than 10 million DALYs. The most prevalent neurological disorders were the various types of headache and Alzheimer's disease and other dementias (table 1).

21Substantial sex differences in age-standardised death, DALY, and prevalence rates existed globally in 2015. The only neurological disorders with less than 10% difference between males and females were Alzheimer's disease and other dementias in terms of death and DALY rates, and meningitis and epilepsy in terms of prevalence rate. Apart from multiple sclerosis, Alzheimer's disease and other dementias, and the three headache types, DALY and prevalence rates were higher in males than females. All these more substantial sex differences were significant apart from those for meningitis, tetanus, and encephalitis death rates and the meningitis and encephalitis DALY rates (appendix pp 153–54). The age pattern varied between the various neurological disorders. The bulk of burden due to communicable neurological disorders occurred in young individuals, particularly the 0–5 years age group. Epilepsy caused the most burden in children and young adults. Headaches peaked at ages 25–49 years, whereas the burden of other neurological disorders increased with age (figure 1).

22Stroke accounted for the largest proportion of total DALYs (47·3%) and deaths (67·3%) among all neurological disorders analysed (figure 2). Migraine, meningitis, and Alzheimer's disease and other dementias were the second, third, and fourth largest contributors of DALYs. The proportional contributions of the other neurological disorders analysed were less substantial and varied from 0·4% (for motor neuron disease) to 5·0% (epilepsy). The second largest contributor to deaths from neurological disorders was Alzheimer's disease and other dementias. The proportional contribution of deaths from other neurological disorders varied from 0·2% for (multiple sclerosis) to 4·0% (for meningitis; figure 2).

23In 2015, the lowest age-standardised DALY rates (less than 3000 per 100 000 people) and death rates (less than 100 per 100 000 people) from neurological disorders were estimated for high-income regions and Latin America, whereas the highest DALY rates (more than 7000 per 100 000 people) were estimated for Afghanistan, Central African Republic, Guinea-Bissau, Kiribati, and Somalia. The highest death rates (more than 280 per 100 000 people) were estimated for Afghanistan and the Central African Republic (figure 3).

24Stroke was the leading cause of age-standardised DALY rates in 18 of 21 GBD regions, while migraine and Alzheimer's disease and other dementias were ranked among the top three causes in all regions except Oceania, south Asia, and the four sub-Saharan African regions, where epilepsy or meningitis ranked higher. Communicable neurological conditions ranked low in high-income regions and central Europe. Epilepsy and medication overuse headache ranked fourth, fifth, or sixth in almost all regions (figure 4).

25A more detailed breakdown of geographical variation is shown as maps of age-standardised prevalence, death rates, and DALY rates for each neurological disorder (appendix pp 156–69). The prevalence of stroke was highest in eastern Europe, central Asia, Oceania, Indonesia, Myanmar, and sub-Saharan African countries. High death rates in Mongolia, Afghanistan, and Central African Republic were responsible for these countries also having the highest age-standardised DALY rates for stroke. Prevalence, death, and DALY rates for Alzheimer's and other dementias were highest in North America and north Africa and the Middle East, while the lowest rates were estimated for sub-Saharan Africa, southern Latin America, and Australasia. Prevalence and DALY rates for Parkinson's disease were highest in high-income regions and lowest in sub-Saharan Africa and eastern Europe. The strong positive relationship with latitude was apparent in the distribution of prevalence, death, and DALY rates for multiple sclerosis. High death rates due to epilepsy were the cause of high DALY rates in sub-Saharan Africa. Prevalence of epilepsy was highest in central Latin America, Chile, north Africa and the Middle East, and Bangladesh. Highest rates of motor neuron disease occurred in high-income regions.

26The prevalence and DALY rates for the different types of headache varied by a factor of three to four between countries with the highest rates in high-income regions, north Africa and the Middle East, and tropical Latin America, while lowest rates were seen in sub-Saharan Africa and east Asia. Medication overuse headache was particularly common in eastern Europe and Iran. The highest death and DALY rates for brain and nervous system cancers were in central Europe and north Africa and the Middle East. Meningitis caused most disease burden in sub-Saharan Africa, especially in the meningitis belt across the Sahara. The largest burden from tetanus was estimated for a few countries in east Africa, particularly South Sudan and Somalia, which have low vaccination coverage owing to national conflict. India had the largest burden from encephalitis.

27Time trends from 1990 to 2015 in prevalence, deaths, and DALYs due to neurological disorders varied across the disorders studied (table 1). Despite a significant drop in age-standardised stroke rates, population increase and ageing combined led to increases in absolute numbers of the three indicators. There was a large increase in the absolute numbers of DALYs, deaths, and prevalent cases of Alzheimer's disease and other dementias even though the changes in age-standardised rates were small. For Parkinson's disease, the numbers of DALYs, deaths, and prevalent cases increased substantially while increases in age-standardised rates were more modest. There was a significant decrease in epilepsy death and DALY rates while the age-standardised prevalence rate remained stable. DALYs, deaths, and prevalent cases of multiple sclerosis increased despite decreasing age-standardised rates. The absolute numbers of DALYs and prevalent cases of motor neuron disease, brain cancer, and headaches increased substantially while there was little change in age-standardised rates. The largest decreases were observed in communicable neurological disorders (tetanus, meningitis, and encephalitis). Although the rates of DALYs and deaths for all neurological disorders combined decreased by more than a quarter from 1990 to 2015, the absolute number of DALYs and deaths from all neurological disorders combined over that period increased by 7·4% (from 233·4 million to 250·7 million) and 36·7% (from 6·9 million to 9·4 million), respectively.

28The decrease in DALY rates from 1990 to 2015 was estimated for most countries. The only countries with an increase in age-standardised DALY rates were the Philippines, Mongolia, Lesotho, Swaziland, and Zimbabwe, largely because these countries had an increase in DALYs from stroke over the 25-year period. The age-standardised DALY rates from neurological disorders dropped by more than half in Ethiopia, Equatorial Guinea, and South Korea (table 2).

29The patterns of disease varied along the development spectrum as measured by the SDI (figure 5). Age-standardised DALY rates of communicable neurological disorders were the largest cause of DALYs at low levels of SDI. Stroke rates increased from low to middle levels of SDI and then decreased to their lowest values at the highest level of SDI. The headaches showed little change in rates with SDI. Rates of epilepsy gradually decreased with rising SDI (figure 5A). However, the all-age rates of combined neurological disorders increased from middle range to highest SDI values, particularly in females (figure 5B). Most of the changes in DALY rates of neurological disorders with development were driven by changes in YLLs. By comparison, the burden due to YLDs showed less variation over the range of SDI (figure 5C).

30On the basis of SDI, Oceania, east and southeast Asia, eastern and central Europe, and central Asia had higher than expected age-standardised DALY rates for stroke for males over the entire estimation period. In females, the stroke DALY rates followed the same pattern but crossed the expected line in a few years of estimation. In females, the stroke DALY rates were initially higher than expected on the basis of SDI but fell to below expected levels during the study period. Latin America, apart from tropical Latin America in the 1990s, eastern and western sub-Saharan Africa, western Europe, Australasia, and high-income North America had lower DALY rates than expected based on SDI (appendix p 170). North Africa and the Middle East had higher than expected DALY rates for Alzheimer's disease and other dementias for males and females. Females in central sub-Saharan Africa, high-income North America, and tropical Latin America had DALY rates close to expected for their SDI values, while all other regions had much lower than expected rates (appendix p 171). Migraine DALY rates in males in south Asia were furthest above the expected line based on SDI. In females, migraine DALY rates in tropical Latin America, south Asia, Australasia, western Europe, and high-income North America were higher than expected. East Asia had much lower DALY rates of migraine than expected in males and females (appendix p 172). Epilepsy DALY rates were higher than expected in eastern, southern, and central sub-Saharan Africa and central Asia. Epilepsy rates were a bit lower than expected in Oceania and east Asia. The DALY rates for meningitis were higher than expected based on SDI in western sub-Saharan Africa. All other regions closely followed the expected pattern (appendix p 173). Plots of regional age-standardised DALY rates and SDI for the remaining neurological disorders are in the appendix (pp 174–83).

31Neurological disorders including stroke, communicable neurological diseases, and brain cancer accounted for 10·2% of global DALYs and 16·8% of all deaths in 2015. DALYs from all neurological disorders combined exceeded those from all injuries (249·8 million), cardiovascular disease (228·9 million, excluding stroke), cancer (209·4 million), and mental and substance use disorders (162·4 million). Our study provides a comprehensive assessment of the extent, patterns, and trends of DALYs for the combined neurological disorders at global, regional, and national levels for 195 countries and territories with important implications for health policy, including priority-setting and financing of health services.

32Despite an overall decrease in the age-standardised rates of DALYs, death, and YLDs of all neurological disorders combined between 1990 and 2015, the number of people dying from and affected by these disorders has increased substantially, contributing to higher health loss across the lifespan. Parkinson's disease was the only neurological disorder with increasing age-standardised rates of deaths, prevalence, and DALYs between 1990 and 2015. Although the burden of communicable neurological disorders has significantly decreased over this period, the burden of non-communicable neurological disorders has significantly increased. This finding is consistent with the overall global burden shift from communicable to non-communicable disorders. In terms of absolute number of people affected by neurological disorders, most of the increase in the burden was associated with ageing of the population and population growth. Increasing incidence of stroke in low-income and middle-income countries, increasing prevalence of multiple sclerosis,, increasing incidence of epilepsy in elderly people, increasing prevalence of tension-type headache, and increasing incidence of brain tumours in elderly people have been reported elsewhere. Findings from other studies have reported that it is difficult to assess trends in prevalence and incidence of Parkinson's disease because of changes in case definitions over time., The evidence on secular trends in prevalence of migraine is mixed.,

33The conclusions from a systematic review of similar studies over time to examine trends in prevalence, incidence, and mortality for people with dementia were that there is some evidence that the incidence of dementia might be declining in high-income countries, but evidence on trends in the prevalence of dementia is inconsistent. Our study showed a modest increase in the prevalence of Alzheimer's disease and other dementias in high-income North America, high-income Asia Pacific, east Asia, south Asia, the Caribbean, and southern sub-Saharan Africa, and a modest decrease elsewhere. The small increase in the global age-standardised prevalence of Alzheimer's disease and other dementias is in contrast with a significant reduction from 1990 to 2015 in the prevalence of stroke, despite stroke sharing risks with, and contributing to, vascular dementia. For stroke, blood pressure control and smoking cessation have been important contributors to the reduction of its incidence in high-income and middle-income countries.,, Diverging trends of stroke and dementia incidence rates have been reported elsewhere. However, the most striking change has been the more than doubling of people in the world who die or are disabled from Alzheimer's disease and other dementias over the past 25 years. As ageing of the global population continues, our findings have important major health-service implications for the care of patients and adequate support for affected families.

34Vaccinations have contributed to the favourable trends in the DALY rates of tetanus, meningitis, and encephalitis.,,,,,, Our estimate of 23·4 million cases of active epilepsy in 2015 is lower than the 32·7 million cases estimated in a meta-analysis of 65 prevalence studies, although that study did not specify a year of estimate. Similar to findings from this meta-analysis, we noted large geographical variations in the prevalence of epilepsy, with significantly greater rates in low-income and middle-income countries. The high prevalence of epilepsy in low-income and middle-income countries can partly be accounted for by the greater number of cases with communicable causes in these countries., Our 2015 prevalence estimates are within the ranges reported for Parkinson's disease (51–177 per 100 000 people), motor neuron disease (1·9–3·9 per 100 000 people),, tension-type headache (21–27%),, medication overuse headache (0·5%–7·2%), and migraine (mean estimate 10%, range 1–25), but lower than that reported for multiple sclerosis (65–74 per 100 000 people), dementias (4·2–8·0% in people aged 60 years or older), and stroke (0·4–2·1% in low-income and middle-income countries). Differences in study populations (eg, age range and countries included) might account for some of the observed differences in prevalence rates.

35Between countries, age-standardised rates of DALYs and deaths from neurological disorders as a group varied up to six times, with the highest rates in low-income to middle-income countries. These geographical patterns of the burden and distribution of individual neurological disorders are important for global and regional health-care planning and might inform further research to examine possible causes of the diseases. For example, the clear latitudinal gradient we noted in the prevalence of multiple sclerosis (about two times higher prevalence in countries at highest latitudes compared with those on the equator) corresponds to that observed in other studies, and is suggestive of the role of environmental factors (eg, vitamin D deficiency and infection).,

36Our findings of large geographical variations in the burden of stroke, dementias, Parkinson's disease, epilepsy, migraine, medication overuse headache, motor neuron disease, and brain and nervous system cancers concur with previous observations.,,,,,,,,,,, In non-communicable neurological disorders, the largest (greater than 20 times) variations in age-standardised DALYs were observed for stroke, motor neuron disease, and multiple sclerosis; in communicable neurological disorders, geographical differences ranged from 100 times for encephalitis to 10 000 times for tetanus. The greater geographical variation of communicable disorders was related to their overwhelming predominance in countries at low levels of socio-demographic development and emphasised the need for better prevention (including vaccination and sanitation measures), as well as better case management in these regions.

37We noted significant sex differences in the burden of many neurological disorders analysed (higher prevalence of tetanus, stroke, Parkinson's disease, and motor neuron disease, but lower prevalence of multiple sclerosis and various types of headaches in males). The greater prevalence of stroke, Parkinson's disease, epilepsy, and motor neuron disease in males has also been reported by other studies. A greater prevalence in females has been reported elsewhere for migraine and multiple sclerosis. Our finding of 22% higher age-standardised prevalence of Alzheimer's disease and other dementias in women is in accordance with a finding from a meta-analysis of consistently higher estimates among 160 studies, although it was reported not to be significant. The higher age-standardised rate might be partly due to having a top age category of 80 years and older. As dementia is so highly prevalent at oldest ages, greater age detail in estimation of dementia prevalence might reduce the observed sex difference.

38Our study findings have important health service implications. The large and increasing numbers of patients with neurological disorders necessitate careful planning by governments and other health-care providers to ensure adequate funding and staff for their treatment and rehabilitation services. However, a recent WHO–World Federation of Neurology survey of services and other resources for neurological disorders in 109 countries (90% of the world population) showed that there are large inequalities in access to neurological care across different populations, in particular for those living in low-income to middle-income countries. With a global shortage of neurologists, neurosurgeons, and rehabilitation professionals, improving neurological care will require innovative strategies within existing health systems. A good example of such innovative strategies is the use of nurses, nurse practitioners, and physician assistants trained in stroke care to care for patients with acute stroke and transient ischaemic attack in stroke units. A recently suggested implementation cycle for combating cardiovascular disease in low-income to middle-income countries provides a good template for similar interventional strategies for reducing the burden of neurological disorders in such countries. Although improving care and rehabilitation of people with neurological disorders is important for improving outcomes, there also are effective primary prevention strategies for communicable neurological disorders and stroke. However, proven effective preventive strategies are often underutilised., More quality epidemiological research on risk factors, incidence, prevalence, and outcomes of neurological disorders in various countries is required to guide better prevention and management of these disorders, and our findings could help to prioritise such efforts. The sex, age, and regional and national differences and trends in the burden of neurological disorders necessitate the development, implementation, and prioritisation of treatments and preventive interventions that are specific for sex, age, and population to reduce the burden from these disorders.

39Although ours was the most up-to-date overview of the global burden of major neurological disorders, this study was not free from some limitations in addition to overall GBD limitations., First, we assumed that the excess mortality in Alzheimer's disease and other dementias implied by prevalence and mortality rates in countries that were most willing to code deaths to dementia in their vital registrations would apply to all other countries and periods. Although we realised that excess mortality was unlikely to be generalisable over location and time, we chose to make this assumption to address the much larger change we observed in deaths certified as Alzheimer's disease and other dementias between countries and over time than seen in prevalence or incidence studies. Second, estimates of cause-specific death rates in most low-income and middle-income countries rely on verbal autopsy data rather than physician-certified death records. Verbal autopsy instruments can identify deaths due to some neurological disorders, including stroke, meningitis, tetanus, and epilepsy, but are unable to capture other neurological disorders. Third, there are also many gaps in data availability by world region, and many low-income and middle-income regions do not have any epidemiological data. Epilepsy is the only neurological disorder with data sources for all 21 GBD regions. Fourth, heterogeneity in study methods and case definitions complicates the non-fatal estimation. Although we endeavoured to adjust for such methodological differences, this relied on generalising adjustment factors from few studies. Fifth, some categories of neurological disease were not included in this analysis because of an inability to aggregate cause-level and sequela-level data. For this reason, we were unable to include secondary epilepsy, the long-term neurological consequences of neonatal disorders, or traumatic brain injury and spinal cord injury, which are estimated in GBD as sequelae of injuries such as falls or road injuries. Adding currently missed neurological disorders would increase the health significance of the neurological disorders for public health systems and should be possible in a future GBD iteration. Sixth, sparse data on the severity of these neurological diseases did not allow us to differentiate severity by location or over time, with the exception of epilepsy (appendix pp 5–113). Relying on few studies, often from high-income countries only, meant we were unable to quantify any treatment effects on severity.

40In conclusion, we have shown that neurological disorders are a large cause of disability and death worldwide. Globally, the burden of neurological disorders has increased substantially over the past 25 years because of population ageing, despite substantial decreases in mortality rates from stroke and communicable neurological disorders. Because low-income and middle-income countries still have a long way to go through the demographic transition of reductions in child mortality and population ageing, the number of patients who will need neurological care will continue to grow in the coming decades. It is important that policy makers and health-care providers are aware of these past trends to be able to provide adequate services for the growing numbers of patients with neurological disorders.

41Correspondence to: Prof Valery L Feigin, National Institute for Stroke and Applied Neurosciences, School of Public Health and Psychosocial Studies, Faculty of Health and Environmental Studies, Auckland University of Technology, North Shore Campus, Auckland 1142, New Zealand valery.feigin@aut.ac.nzandProf Theo Vos, Institute for Health Metrics and Evaluation, University of Washington, Seattle, WA 98121, USA tvos@uw.edu

42Correspondence to: Prof Valery L Feigin, National Institute for Stroke and Applied Neurosciences, School of Public Health and Psychosocial Studies, Faculty of Health and Environmental Studies, Auckland University of Technology, North Shore Campus, Auckland 1142, New Zealand valery.feigin@aut.ac.nz

43and

44Prof Theo Vos, Institute for Health Metrics and Evaluation, University of Washington, Seattle, WA 98121, USA tvos@uw.edu

45For more on the Global Health Data Exchange see http://ghdx.healthdata.org/gbd-2015/data-input-sources

46For more on the Global Health Data Exchange see http://ghdx.healthdata.org/gbd-2015/data-input-sources