Chapter 39 Neuroepidemiology
Population-Based Rates
The incidence or attack rate is the number of new cases over a defined study period divided by the population at risk. This usually is given as an annual incidence rate in cases per 100,000 population per year. The date of onset of clinical symptoms typically dictates the time of accession, although occasionally the date of first diagnosis is used. The (point) prevalence rate is more properly called a ratio, but it refers to the number of those affected, both old and new cases, at one point in time within the community per unit of population. The lifetime prevalence rate refers to the proportion of persons manifesting a disorder of interest during the period of their life up to the survey date. It typically is reported per 1000 of the population at risk. If no change in case-fatality ratios occurs over time and no change in annual incidence rates (and no migration) occurs, then the average annual incidence rate times the average duration of illness in years equals the point prevalence rate. When numerator and denominator for a rate each refer to an entire community, their quotient is a crude rate, for all ages. When both terms of the ratio are delimited by age or sex, these are age-specific or sex-specific rates, respectively. Such rates for consecutive age groups, from birth to the oldest group of each sex, provide the best description of a disease within a community. In comparing morbidity or mortality rates between two communities for an age-related disorder (such as stroke or epilepsy), differences in crude rates may be observed solely because of differences in the age distributions of the denominator populations. This can be avoided by comparing only the individual age-specific rates between the two, but this approach rapidly becomes unwieldy. Methods exist for adjusting the crude rates for all ages to permit such comparisons. One such method involves taking community age-specific rates and multiplying them by the proportion of a “standard” population within the same age group. The sum of all such products provides an age-adjusted (to a standard) rate, or a rate for all ages adjusted to a standard population. One common standard in the United States is its population for a given census year. The mortality or death rate is the number of deaths in a population in a period with a particular disease as the underlying cause, such as an annual death rate per 100,000 population. Deaths by cause are provided by official government agencies, based on standard death certificates. At times, deaths listed as other than underlying cause on the certificate are added to give a count of total deaths for the disease. The standardized mortality ratio (SMR) is the observed number of deaths in the study group of interest divided by the expected number of deaths based on the standard population rates applied to the study group. The great advantage of death rates is their current availability over time and geographical area for many disorders, whereas morbidity rates require specific community surveys. Geographical distributions from death data are especially informative because most population studies available are, of necessity, spot surveys that may tell little about areas that were not investigated. Most often, too, the numbers are larger by orders of magnitude than those that prevalence studies can provide. The principal disadvantage, and it is a major one, is the question of diagnostic accuracy. In clinical practice, the diagnostic code used for mortality rates is a three- or four-digit number representing a specific diagnosis in the International Statistical Classification of Diseases, Injuries and Causes of Death (ICD), which is revised periodically. The changes in the 10th revision (ICD-10) were major. ICD-10 was published in 1992 and adopted for use in the United States in 1999. It introduced the innovation of an alphanumerical coding scheme of one letter followed by three numbers (e.g., I63.1, cerebral infarction due to thrombosis of precerebral arteries). One drawback of the ICD system of classification is that different diseases frequently are subserved under the same primary code. To provide a more refined classification for individual diseases, several disciplines have published specialty-related expansions of the primary ICD structure. ICD-10-NA is the expansion of the codes relating to neurological diseases, so that virtually every known neurological disease or condition has a unique alphanumerical identifier (van Drimmelen-Krabbe et al., 1998; World Health Organization, 1997). In the United States, the Department of Health and Human Services has mandated conversion to ICD-10 for all healthcare organizations by 2013. Lack of space here precludes attention to community survey methods, risk factors and analytic epidemiology, treatment comparisons, and statistical methods—all intrinsic aspects of epidemiology. This chapter highlights the descriptive epidemiological analysis for a few major neurological diseases selected as representative of those most likely to be encountered in clinical neurology.
Cerebrovascular Disease
Stroke (see Chapter 51) is the third leading cause of death and a major cause of disability in the United States (Miniño et al., 2009). Most recent epidemiological studies have subdivided stroke into subarachnoid hemorrhage (SAH), intracerebral hemorrhage, and cerebral infarction. Subdural hemorrhage is not included in this category. Cerebral infarction is the most common type of stroke in developed countries, making up more than 70% of cases. Intracerebral hemorrhages account for approximately 10% to 15% of strokes, and SAHs make up less than 5%, while the remainder are of undetermined etiology.
Mortality Rates
Since the late 1960s, U.S. stroke death rates have declined by 60% overall (Howard et al., 2001). The largest declines in stroke mortality were seen in white men and the smallest in black men. Similar decreasing rates in stroke mortality are reported for other countries including Japan, Australia, New Zealand, Canada, and all of Western Europe. Recent annual age-standardized mortality rates in Europe ranged between 26 and 50 per 100,000 for men and between 18 and 23 per 100,000 for women. Reported mortality rates over recent decades have actually increased for Eastern Europe (Sarti et al., 2003).
The geographic differences in stroke mortality within the United States are notable, with the highest rates in the southeastern region since the 1940s. The so-called stroke belt states of Georgia, North Carolina, and South Carolina have consistently demonstrated mortality rates above the U.S. average. The Reasons for Geographic and Racial Difference in Stroke (REGARDS) Study was launched in 2001 to better understand demographic differences in stroke. Recent data from this project have shown that geographic differences in risk factors contribute little to explaining geographic variations in stroke mortality (Howard et al., 2009).
In general, age-specific death rates for stroke exhibit a logarithmic increase with increasing age. Racial and ethnic disparities in mortality from stroke have been recorded in many studies in the United States. The Centers for Disease Control and Prevention (CDC) examined this issue by evaluating 2002 U.S. mortality and death certificate data. During 2002, 12% of all stroke deaths occurred among persons younger than 65 years of age (Centers for Disease Control and Prevention [CDC], 2005). Age-specific death rates were notably higher in blacks than in whites by factors of 2.5, 3.5, 2.8, and 1.9 for the successive age groups 0 to 44, 45 to 54, 55 to 64, and 65 to 74 years. For all other groups (Asian/Pacific Islanders, American Indians/Native Americans, Hispanics), little consistent difference from the whites was observed in any age group (Fig. 39.1). United States racial and ethnic disparities in mortality by stroke type have been reported for the period 1995 to 1998 (Ayala et al., 2001). National Vital Statistics death certificate data were used to calculate age-standardized rates for ischemic stroke, intracerebral hemorrhage, and SAH among Hispanics, blacks, American Indians/Alaska Natives, Asians/Pacific Islanders, and whites. For ischemic stroke, the ratio of the age-standardized death rate among blacks (96 per 100,000) compared with whites was 1.30; all other groups had lower rates than the whites. Death rates for intracerebral hemorrhage were highest for blacks (23 per 100,000) and Asian/Pacific Islanders (20 per 100,000), with corresponding risk ratios compared with whites of 1.70 and 1.52. Subarachnoid hemorrhage death rates were higher in whites than for all minority groups.
Morbidity Rates
Like mortality rates, stroke incidence has declined rapidly over the past 50 years. Within the past 2 decades, however, the incidence rates have seemed to plateau or decrease only slightly in industrialized countries (Kleindorfer et al., 2010). Stroke incidence increases logarithmically with increasing age but with a lesser increase beyond age 74. Average annual age-adjusted incidence rates by sex show a modest, but possibly increasing, male excess. In recent years, the annual stroke incidence rate in Europe and North America has been between 100 and 300 per 100,000 population, and mostly near 150.
The most recent trends in stroke incidence were reported from population-based stroke registries in six European countries (Heuschmann et al., 2009). Total stroke incidence for any first stroke between 2004 and 2006 and adjusted to the European population was 141 per 100,000 for men and 95 per 100,000 for women. Stroke incidence ranged in men from 101 per 100,000 in Sesto Fiorentino to 239 per 100,000 in Kaunas; and in women, from 63 per 100,000 in Sesto Fiorentino to 159 per 100,000 in Kaunas. The median age for first stroke was 73 years, with 51% occurring in females. On average, the highest rates were observed in Eastern Europe and the lowest in Southern Europe. Incidence rates in the United States for blacks remain higher than those for whites. The Greater Cincinnati and Northern Kentucky Stroke Study was the first large metropolitan-based study of stroke trends among blacks (Kleindorfer et al., 2010). The incidence for stroke between 1993 and 2005 decreased significantly for whites but was stable for blacks. These changes were driven by a drop in ischemic stroke among whites but stable ischemic stroke rates among blacks. Case fatality ratios did not differ by race. In the United States, age-, race-, and sex-adjusted stroke prevalence rates increased from 1.41% in the period 1971 to 1975 to 1.87% in the period 1988 to 1994 (Muntner et al., 2002). This corresponded with an increase of 930,000 noninstitutionalized stroke survivors, with increases observed in all age, race, and gender groups. With decreasing mortality trends and relatively stable stroke incidence rates during the 1980s, these data point to a decreasing stroke case-fatality ratio as a major reason for the increasing prevalence.
Transient Ischemic Attacks
Although clearly a subset of cerebrovascular disease, transient ischemic attacks (TIAs) generally have been excluded from most morbidity and mortality surveys of stroke. As with stroke incidence and prevalence rates, a marked increase in TIA rates occurs with age. The Oxford Vascular Project found a slight increase in the incidence of TIAs between 1981 and 1984 and between 2002 and 2004, with overall rates rising from 0.33 per 1000 to 0.51 per 1000 (Rothwell et al., 2004). TIAs in persons older than 65 years of age accounted for the major part of this rate increase between time periods. This trend was also confirmed in a community-based study of older adults in Korea, where an age- and education-adjusted prevalence of TIA of 8.9% (age 65+) was found (Han et al., 2009).
The new tissue-based definition for TIA takes into account recent neuroimaging findings, as well as providing a much shorter duration for the diagnosis (Albers et al., 2002). If the new definition were to be used in epidemiological studies, the estimated annual incidence of TIA would be lowered by 33% and the incidence of ischemic stroke increased by 7% (Ovbiagele et al., 2003). However, a major underascertainment of TIA is probable in all surveys, with undefined differences among them. This also may give spuriously high frequencies for completed stroke after TIA because in many studies of stroke, only a retrospective history of TIA occurrence is given.
Primary Neoplasms
Three large centralized U.S. databases have been created that provide descriptive epidemiological data on primary brain tumors (see Chapter 52A). These databases include the Central Brain Tumor Registry of the United States (CBTRUS); the Surveillance, Epidemiology and End Results (SEER) database; and the National Cancer Database (NCDB). According to the CBTRUS database, a total of 158,088 persons were diagnosed with a primary brain or central nervous system (CNS) tumor in the United States in the years 2004-2006 (CBTRUS, 2010). The lifetime risk of developing a CNS tumor is estimated to be 0.65% for men and 0.50% for women (Ries et al., 2005). Little is known of the causes of most primary brain tumors, but their epidemiological features may provide clues for more definitive studies.
Mortality Rates and Survival
In the United States for 1995-1999, malignant CNS tumor deaths by age showed a steep rise from very low rates in early adult life to a peak of approximately 20 per 100,000 per year by age 75, followed by a steep decline with further increasing age (Davis et al., 2001). These rates presumably are chiefly for glioblastoma multiforme. A notable excess of whites over nonwhites was seen in this group, with rates two to three times higher in the white patients. An excess of male deaths occurred in all racial groups.
Reported 5-year survival ratios have been approximately 60% for clinically diagnosed meningioma and 20% for gliomas as a group. When these two tumor types are taken together, median survival for benign brain tumors may be estimated at 6 years. The relative 5-year survival rate for children younger than 15 years of age with brain and other nervous system tumors is now 61%, compared with 35% some 20 years ago (Parker et al., 1997). Population-based data between 1990 and 2001 from the United Kingdom showed that 5-year survival rates for all CNS tumors for those 15 to 29 years of age were slightly worse than for those 0 to 14 years of age (62% versus 67%) (Feltbower et al., 2004). Glioblastoma is the most common primary brain tumor in adults, with a uniformly poor prognosis. Median survival for glioblastoma remains approximately 1 year after diagnosis. Several studies from cancer registries have indicated that the 5-year survival rate, typically reported at 4% to 10% over the past 3 decades, may be too optimistic (Tran and Rosenthal, 2010). Series from Canada, Sweden, and the United States that reviewed clinical and histological data from registries found that in half of all reported cases of glioblastoma, the tumor had been misclassified and on close inspection was found to be a less aggressive tumor (McLendon and Halperin, 2003). Corrected 5-year survival rates are more likely to be in the 2% to 3% range. Some positive news for a subgroup of glioblastoma patients with the MGMT (O6-methylguanine-DNA methyltransferase) DNA repair gene was recently reported (Hegi et al., 2005). Irrespective of treatment, patients with glioblastoma with a methylated MGMT promoter survived approximately 55% longer than patients with an unmethylated MGMT promoter. The gain, although real, was therefore only some 6 months. The methylation of the MGMT promoter gene compromises DNA repair and triggers cytotoxicity. In addition, patients with glioblastoma and the MGMT promoter also demonstrated an improved treatment response to alkylating chemotherapy agents. The epidemiology of metastatic brain tumors is that of the primary cancer. Survival for patients with metastatic brain tumors is poor. Even after whole-brain irradiation, median survival is approximately 6 months (Andrews et al., 2004). Adjuvant therapy with sterotactic radiosurgery boost may extend survival for patients with a small number of metastases. Survival for 740 patients with brain metastases was reviewed by Hall and colleagues (2000). For all tumor types, the actuarial survival rate was 8.1% at 2 years, 4.8% at 3 years, and 2.4% at 5 years. At 2 years from diagnosis, ovarian carcinoma had the highest survival rate (23.9%) and small cell lung cancer (SCLC) the lowest (1.7%). Favorable prognostic variables for survival included a single metastatic lesion, surgical resection, and whole-brain irradiation.
Morbidity Rates
Using the SEER database, Gurney and Kadan-Lottick (2001) calculated incidence trends by age group for malignant brain tumors for 1975 to 1997 (Fig. 39.2). Incidence rates remained stable for persons 20 to 69 years of age during the period. A 35% increase in rates for children 0 to 14 years old was seen in the mid-1980s. A gradual increase in malignant tumor rates was observed for persons older than 70 years between 1975 and 1990. During the 1990s, the rates for most groups remained essentially stable. Increasing incidence trends must be interpreted with caution. At least some of these changes can be attributed to the dramatic improvements in neuroimaging seen from the 1980s on. In meningioma, age-specific rates continue to rise with age to the oldest group, and a female preponderance is found. The suspected excess in blacks was borne out in a survey in the Los Angeles County Cancer Surveillance program. Age-adjusted average annual incidence rates for meningiomas were 1.8 per 100,000 males and 2.7 per 100,000 females. Respective non-Hispanic white rates were 1.8 per 100,000 and 2.5 per 100,000; for blacks, they were 2.5 per 100,000 and 3.6 per 100,000. In Rochester, Minnesota, annual incidence rates were 4.9 per 100,000 males and 5.8 per 100,000 females for the years 1935 to 1977, but only 1.2 and 2.6 per 100,000 respectively for cases diagnosed before death.
The most recent overall incidence estimate for malignant intracranial tumors in the United States is 7.2 per 100,000 person-years population for 2004-2006 (CBTRUS, 2010). For benign brain tumors for the same period, the figure is 11.5 per 100,000 person-years, including 6.3 per 100,000 for meningiomas and 2.5 for tumors of the sella region.
Metastatic brain tumors are more common than primary malignant brain tumors, with incidence rates of approximately 10 per 100,000. The relative frequencies of brain metastases, called incidence proportions (IPs), in patients diagnosed in the Metropolitan Detroit Cancer Surveillance System between 1972 and 2001 were reported by Barnholtz-Sloan and associates (2004). Total IP of brain metastases was 9.6% for all primary sites combined, with highest IPs for lung (19.9%), melanoma (6.9%), renal (6.5%), breast (5.1%), and colorectal (1.8%) cancers. African Americans demonstrated higher IPs than other racial groups. This total IP is lower than in earlier reports, which had ranged from 20% to 50%.
Although some CNS tumors have a clear genetic character, less than 5% can be attributed to inheritance. Many risk factors have been implicated in human brain tumors, the vast majority of which are unsubstantiated by scientific evidence. High-dose irradiation leads to an increased incidence of primary brain tumors, but the association of higher brain tumor risk with low doses of radiation is more controversial. Prolonged cell phone use and risk for brain tumors has been the subject of several studies over the past decade, with mixed results (Ahlbom et al., 2009). Overall, cell phone use studies to date do not demonstrate an increased risk of brain tumors within 10 years. For slow-growing tumors such as meningiomas or acoustic neuromas, the absence of an association is less conclusive with such a limited observation period.
Recent epidemiological studies suggest prenatal and early childhood environmental factors that may alter the risk of brain tumors. Increased risk for brain tumors was identified in persons born in late fall through early spring in one report (Brenner et al., 2004) and has been associated with maternal smoking during pregnancy in another (Brooks et al., 2004). The protective effects of vitamin supplementation during pregnancy have been borne out in several studies (Preston-Martin et al., 1998).
Convulsive Disorders
Epilepsy is defined as recurrent seizures (i.e., two or more distinct seizure episodes) that are unprovoked by any immediate cause (see Chapter 67). The International League Against Epilepsy (ILAE) classification system divides the epilepsies into four broad groups: (1) localization-related; (2) generalized; (3) undetermined whether localized or generalized; and (4) special syndromes (Everitt and Sander, 1999). Within the localization-related and generalized groups, further subdivisions into symptomatic (known cause), idiopathic (presumed genetic origin), and cryptogenic (no clear cause) are recognized. The major clinical types of seizures are generalized tonic-clonic, absence, incomplete convulsive (myoclonic), simple partial (focal), and complex partial (temporal lobe or psychomotor). Status epilepticus is defined as any seizure lasting for 30 minutes or longer, or recurrent seizures for more than 30 minutes during which the patient does not regain consciousness.
Mortality Rates
Reported mortality rates with epilepsy are on average two to three times greater than those in the general population. Shackleton and colleagues performed a meta-analysis on 21 studies of epilepsy mortality and found overall SMRs between 1.2 and 9.3 (Shackleton et al., 2002). Population-based studies with long-term follow-up give SMRs between 2 and 4, which seem the more accurate estimates.
As to evaluating cause of death, the proportionate mortality (PMR) is frequently used. The PMR for conditions related to epilepsy range between 1% and 13% for population-based studies (Hitiris et al, 2007). Etiologies include status epilepticus and seizure-related causes (PMR 0% to 10%), sudden unexplained death in epilepsy (SUDEP; PMR 0% to 4%), suicide (PMR 0% to 7%), and accidents (0% to 12%). Causes of nonepilepsy-related death include ischemic heart disease (PMR 12% to37%), cerebrovascular disease (PMR 12% to 17%), cancer (PMR 18% to 40%), pneumonia (PMR 0% to 7%), suicides (PMR 0% to 12%), and accidents (0% to 4%).
SUDEP generally is considered to be the most common cause of epilepsy-related death, with a relative frequency of 1 per 1000 epilepsy cases (Opeskin and Berkovic, 2003). Risk factors that have been consistent across studies include male sex, generalized tonic-clonic seizures, early age of onset of seizures, refractory treatment, and being in bed at the time of death. Proposed mechanisms for SUDEP include central apnea, acute neurogenic pulmonary edema, and cardiac arrhythmia precipitated by seizure discharges acting via the autonomic nervous system. Other causes of death in epilepsy can be classified as those in which epilepsy is secondary to an underlying disease (cerebrovascular disease) or is an unrelated disorder (ischemic heart disease). Age-specific mortality rates for Rochester, Minnesota, are shown in Fig. 39.3. Graphed curves for mortality data were similar in configuration to those for age-specific prevalence data, but rates were 1000-fold lower. This finding suggests that each year, 0.1% of the patients with epilepsy die of causes directly related to their epilepsy. Status epilepticus affects 105,000 to 152,000 persons annually in the United States (DeLorenzo et al., 1996). Status epilepticus represents a neurological emergency, and despite improvements in treatment, the mortality rate is still high. Population-based studies have reported 30-day case-fatality ratios between 8% and 22%. Short-term fatality after status epilepticus is associated with the presence of an underlying acute etiological disorder. Fatality ratios are lowest in children (short-term mortality rate 3% to 9%) and highest in the elderly (short-term mortality rate 22% to 38%). Case-fatality ratios for those surviving the initial 30 days after status epilepticus are 40% within the next 10 years.
Morbidity Rates
Fig. 39.3 also shows morbidity measures for epilepsy in Rochester, Minnesota, by age group. Age-specific incidence of epilepsy was high during the first year of life, declined during childhood and adolescence, and then increased again after age 55. The cumulative incidence of epilepsy was 1.2% through age 24 and steadily increased to 4.4% through age 85 years. Age-specific prevalence increased with advancing age; nearly 1.5% of the population older than 75 years had active epilepsy.
Point prevalence and average age-adjusted annual incidence rates for epilepsy are available from a number of community surveys (Banerjee et al., 2009
