Survival and epidemiology of amyotrophic lateral sclerosis (ALS) cases in the Chicago and Detroit metropolitan cohort: incident cases 2009–2011 and survival through 2018

Abstract

 

Amyotrophic lateral sclerosis (ALS) is a fatal, progressive neurodegenerative disorder. The National ALS Registry launched surveillance projects to understand the distribution of ALS in targeted geographic cohorts.

Objective

To describe the demographics, incidence, and survival of persons with ALS (PALS) identified in the Chicago and Detroit area population-based cohort.

Methods

Neurologists in the catchment area provided case reports for eligible ALS cases diagnosed and/or cared for from 1 January 2009 through 31 December 2011. Crude incidence rates were calculated for 2009–2011 and stratified by race and ethnicity. Using data from the National Death Index through 2018, we modeled the effect of patient covariates on mortality using the Cox proportional hazard regression.

Results

Of the 574 cases, 372 (64.8%) were diagnosed from 2009 to 2011. The combined crude incidence rates for 2009, 2010, and 2011 were 1.44, 1.53, and 1.73 cases per 100,000 person-years, respectively. Of the 486 subjects with complete survival data, 81% were deceased at the end of follow-up. Median survival time was 2.2 years, with 30% and 9% of subjects surviving past 5 and 10 years after diagnosis, respectively. Additionally, female PALS and PALS with longer time between symptom onset and diagnosis experienced longer survival. Nonwhites also experienced longer survival than Whites, except for those cases diagnosed in the younger age categories.

Conclusion

Understanding the survival of ALS patients can aid in understanding variable prognostic factors, which can potentially extend survival and improve disease management.

Keywords:

• Amyotrophic lateral sclerosis (ALS)
• surveillance
• incidence
• survival
• hazard ratio

Introduction

Amyotrophic lateral sclerosis (ALS) is a devastating, neurodegenerative disorder affecting motor and extra-motor system neurons. Upper and lower motor neuronal death results in fatal, progressive limb, and bulbar disability (1). Moreover, extra-motor system neuronal loss may impart cognitive and behavioral symptoms (2). ALS was first characterized by Charcot and Joffroy in 1869 and there is only a modest understanding of pathogenesis and prognostic determinants (3,4). Excluding familial ALS, accounting for 5–10% of cases, there are suspected but no known causes of ALS. Suspected risk factors of sporadic ALS include smoking, athletic activity, history of military service, head trauma, heavy metal exposure, toxic chemical exposure, and exposure to electromagnetic fields (5). Prognosis can vary drastically; estimated median survival from symptom onset is approximately 2–4 years, with only 5–10% of patients living 10 years after the start of their initial symptoms (6). Commonly reported predictors of worse prognosis are older age at diagnosis and bulbar onset. Moreover, diagnostic delay and longer time to diagnosis from symptom onset has been associated with longer survival (6).

The National ALS Registry (Registry) was created in 2008 in response to the ALS Registry Act and is administered by the federal Agency for Toxic Substances and Disease Registry (ATSDR). The registry collects, disseminates, and analyzes data on ALS cases across the United States (U.S.) and their potential risk factors. Recent registry reports conservatively estimate the prevalence of ALS cases for the U.S. was 5.20 per 100,000 persons in 2016 (7). This is comparable to global and European estimates of ALS median prevalence of 4.48 per 100,000 and 5.40 per 100,000, respectively (8). As ALS is not a nationally notifiable disease in the U.S., collecting incident data on all ALS cases is a challenge. Registry efforts were supplemented by surveillance projects collecting incident and prevalence data in three states and eight metropolitan areas selected to overrepresent minorities. All surveillance projects combined reported ALS incidence estimates in the range of 1.40–1.90 per 100,000 persons (9–13). Similarly, these estimates are comparable to estimates of pooled global incidence of 1.68 per 100,000 persons (14).

This analysis describes the demographics, incidence, and survival of persons with ALS (PALS) identified in the Registry’s Detroit and Chicago Metropolitan Area ALS Surveillance Projects. Understanding the demographics of those who get ALS and estimating incidence may aid in better resource allocation. Moreover, characterizing the survival experiences of PALS can aid in understanding mutable prognostic factors that can extend life and improve disease management.

Materials and methods

Analysis population

The ALS cases used for this analysis were collected as part of ATSDR’s State and Metropolitan Area ALS Surveillance Project. Through a partnership with the Detroit Department of Health and Wellness Promotion (DHWP) and the Metropolitan Chicago Healthcare Council (MCHC), neurologists were contacted to determine if they diagnosed and/or cared for ALS patients. A person with ALS was eligible to be reported if diagnosed and/or cared for from 1 January 2009 through 31 December 2011, and was a resident of one of the project’s catchment areas. The Chicago catchment area covered Cook (includes the city of Chicago) and DuPage counties, and the Detroit catchment area included only residents of Wayne County. A higher emphasis was placed on neurologists who specialized in ALS case and typically saw more than 50 patients per year. Neurologists were requested to complete a one-page Case Reporting Form (CRF) which included patient demographic questions such as sex, date of birth, race, ethnicity, date of diagnosis, El Escorial criteria, date of onset of symptoms, family history, provider details, and payer information. Data were retrospectively collected from the neurologists for all eligible prevalent and incident cases. State specific mortality data were also reviewed to identify additional missed ALS patients. A detailed description of the methodology used for case ascertainment can be found in the state and metropolitan area-based ALS surveillance paper by Wagner et al. (15).

The rationale behind combining the analysis of the Chicago and Detroit ALS State and Metro surveillance projects was due to their geographic and demographic similarities. Chicago and Detroit were two of the largest metropolitan areas located in the Midwest. The Chicago area had a 30.1% African American population, where Detroit had a slightly higher (38.8%) African American population compared with 12.6% for the U.S. (16). The data collection methodologies were the same for both projects.

Statistical methods

In this analysis, several baseline characteristics of ALS cases in the Chicago and Detroit metropolitan areas, were tabulated including age at diagnosis, sex, race, ethnicity, El Escorial ALS case criteria classification, provider type, history of dementia, and family history of ALS. Subjects with unidentified birth year were excluded from analysis. Date of diagnosis was defined as the earliest date that ALS was confirmed by a general neurologist. The majority of reported cases were confirmed by a neurologist specializing in ALS. Age at diagnosis was categorized into six groups: 18–39 years, 40–49 years, 50–59 years, 60–69 years, 70–79 years, and 80 years or more, in correspondence to the categorization in the yearly ALS prevalence report from the National ALS Registry. Race was categorized as “White”, “African American”, “Asian”, and “Unknown”.

Expected number of cases for both cohorts were determined using the 2010 U.S. Census population data and estimates of the national incidence and prevalence. Crude incidence rates were calculated using the count of cases diagnosed in each year as the numerator and the corresponding U.S. Census population data as the denominator. Crude average annual incidence rates were calculated by adding the incidence rates for the three years and then dividing by three. All incidence rates were adjusted to 100,000 person-years (PY). Incidence rates were also stratified by race and ethnicity for both the individual Chicago and Detroit cohorts.

Cases were submitted to the National Death Index (NDI) to obtain vital status through 31 December 2018 and matched to the Chicago and Detroit cases by unique case identification numbers. We used Cox proportional hazards regression to model the effect of patient covariates on mortality (17). In all modeling, date of ALS diagnosis was the time origin and time on study the time axis. Subjects not determined to be deceased via the NDI were censored on 31 December 2018. Subject matter knowledge of the suspected and known sources of confounding guided covariate selection. We estimated hazard ratios (HRs) with 95% confidence intervals (CI) for the following covariates: sex, age at diagnosis (years, continuous), race/ethnicity (non-Hispanic White vs. other race or Hispanic), diagnosis type (definite or probable ALS vs. possible ALS or not classifiable), and time interval between symptom onset and diagnosis (years, continuous). Additionally, race/ethnicity was categorized as “White” to include white race and non-Hispanic ethnicity and “non-White” to be all minority non-white races and Hispanic ethnicity. In all models, incomplete records for patient covariates (i.e. listwise deletion) we excluded. A p value of less than 0.05 was deemed statistically significant. All analyses were conducted with SAS 9.4 (SAS Institute, Inc., Cary, NC). Data for this project were collected under a protocol approved by the Centers for Disease Control and Prevention Institutional Review Board.

Results

Demographic characteristics

The baseline demographic characteristics of all the incident and prevalent ALS cases identified in the Chicago and Detroit metropolitan area are described in Table 1. Based on the 2010 U.S. Census population and estimates of incidence and prevalence, the project identified 88% (431/489) of the expected cases in Chicago and 99% (144/146) of the expected cases in Detroit. In the combined cohort, the overall average age at diagnosis was 61 years, and a large portion of cases were diagnosed between 50 and 79 years of age. In Chicago, the largest percentage of reported cases were in the 60–69 age group, while Detroit cases were in the 50–59 age group. Thus, reported cases in Detroit generally had a younger age at diagnosis, compared with cases in Chicago. In both cities, more males than females were reported with ALS. Of the 574 total cases, 81 (14.1%) were African American, with 49 (11.4%) cases in Chicago and 32 (22.2%) cases in Detroit. This percentage of African American cases was higher than that from the overall State and Metropolitan Area ALS surveillance project.

Table 1 Demographic characteristics of reported ALS cases in the Chicago and Detroit metropolitan area, 1 January 2009 to 31 December 2011 (n = 574).

		Combined		Chicago		Detroit
		Count of cases	%	Count of cases	%	Count of cases	%
Age at diagnosis	18–39	34	5.9	27	6.3	7	4.9
	40–49	72	12.5	56	13	16	11.1
	50–59	155	27.0	107	24.9	48	33.3
	60–69	158	27.5	123	28.6	35	24.3
	70–79	115	20.0	86	20.7	26	18.1
	≥80	40	7.1	28	6.5	12	8.3
Gender	Male	304	53	220	51.2	84	58.3
	Female	270	47	210	48.8	60	41.7
Race	Asian	17	3	16	3.7	1	0.7
	African American	81	14.1	49	11.4	32	22.2
	White	394	68.6	285	66.3	109	75.7
	Unknown	82	14.3	80	18.6	2	1.4
Ethnicity	Hispanic	46	8	44	10.2	2	1.4
	Non-Hispanic	454	79.1	373	86.8	81	56.2
	Unknown	74	12.9	13	3	61	42.4
El Escorial criteria	Definite	173	30.1	117	27.2	56	38.9
	Probable	174	30.3	126	29.3	48	33.3
	Probable and Lab supported	89	15.5	68	15.8	21	14.6
	Possible	78	13.6	61	14.2	17	11.8
	Unclassified	60	10.5	58	13.5	2	1.4
Provider	ALS specialist	552	96.2	414	96.3	138	95.8
	Other	22	3.8	16	3.7	6	4.2
Dementia	Yes	38	6.6	26	6	12	8.3
	No	412	71.8	316	73.5	96	66.7
	Unknown	124	21.6	88	20.5	36	25
Family history	Yes	26	4.5	20	4.6	6	4.2
	No	483	84.2	380	88.4	103	71.5
	Unknown	65	11.3	30	7	35	24.3
Total		574	100	430	74.9	144	25.1

Incidence

During 2009–2011, the Metropolitan Area ALS Project identified 430 cases in Chicago and 144 cases in Detroit. Of the 574 combined cases in Chicago and Detroit, 372 (64.8%) were diagnosed during 2009–2011. The crude incidence rates for the combined Chicago and Detroit area for 2009, 2010, and 2011 were 1.44, 1.53, and 1.73 cases per 100,000 PY, respectively. The average annual incidence rate for the three-year period was 1.56 cases per 100,000 PY (Table 2).

Table 2 Annual incidence rates for ALS cases diagnosed in a 3-year period in the Chicago and Detroit metropolitan areas by race and ethnicity, 2009–2011 (n = 372).

	2009		2010		2011		2009–2011
	Cases	Incidence	Cases	Incidence	Cases	Incidence	Total cases	Average incidence per 100,000 person years
Chicago (total)	88	1.44	87	1.42	103	1.69	278	1.52
Race^a
White	61	1.48	53	1.28	72	1.74	186	1.50
African American	8	0.62	12	0.93	7	0.54	27	0.70
Asian	2	0.38	4	0.76	3	0.57	9	0.57
Ethnicity^a
Hispanic	12	0.82	14	0.96	7	0.48	33	0.75
Non-Hispanic	76	1.64	72	1.55	91	1.96	239	1.71
Detroit (total)	26	1.43	34	1.87	34	1.87	94	1.72
Race^a
White	21	2.11	26	2.62	25	2.51	72	2.41
African American	5	0.71	6	0.85	8	1.14	19	0.90
Asian	0	0.00	1	1.57	0	0.00	1	0.52
Ethnicity^a
Hispanic	0	0	0	0	0	0	0	0
Non-Hispanic	16	0.94	21	1.23	20	1.17	57	1.11
Combined	114	1.44	121	1.53	137	1.73	372	1.56

^a Unknown and missing race and ethnicity are excluded.

Of the 430 reported cases in Chicago, 278 (64.7%) were diagnosed during 2009–2011. Of the 144 reported cases in Detroit, 94 (65.3%) were diagnosed during 2009–2011. The annual incidence rate for the three-year period was slightly higher in Detroit than that in Chicago (Table 2). Similarly, a higher incidence rate was reported for the African American population in the Detroit cohort (0.90) compared with the Chicago cohort (0.70 cases per 100,000 PY).

Survival analysis

Among the 486 analysis subjects with complete patient data, 393 (81%) were deceased at the end of follow-up. Median survival time was 2.2 years (95% CI 2.1, 2.7) with 30% and 9% of subjects surviving past 5 and 10 years after diagnosis, respectively. Figure 1 shows the plot of the survivor function. Table 3 shows the fitted Cox model. Except for sex, all covariates were statistically significant (p < 0.05) in addition to an interaction between race/ethnicity and age at diagnosis. Table 4 shows HRs for selected covariates. Female PALS (HR 0.91; 95% CI 0.74, 1.12) and PALS with longer time between symptom onset and diagnosis (HR 0.83; 95% CI 0.77, 0.89) experienced longer survival. Conversely, patients with a definite or probable ALS diagnosis (HR 1.26; 95% CI 1.06, 1.73) and those who received an ALS diagnosis later in life experienced shorter survival. It is important to note that the later survival estimates varied with race/ethnicity. A one-year increase in age at diagnosis resulted in a 5% greater hazard among non-Hispanic Whites compared with a 2% increase among non-Whites or Hispanics (HR 1.05 [95% CI 1.04, 1.06] and 1.02 [95% CI 1.01, 1.04], respectively). Figure 2 illustrates differences in survival experiences for PALS receiving their diagnosis at different ages by race and indicates among patients receiving their ALS diagnosis at 70 years of age or older experienced significantly worse survival for non-Hispanic Whites after one year after diagnosis.

Figure 1 Plot of the survivor function for the Chicago and Detroit metropolitan area cohort.

Figure 2 Survivor functions for non-Hispanic Whites and Hispanic or nonwhite cases at ages 40 (a), 55 (b), and 70 years at diagnosis (c). Other covariates were fixed at male sex, definite or probable ALS diagnosis and a one-year interval between symptom onset and diagnosis.

Table 3 Cox model fit to the Chicago/Detroit Surveillance Data, 18 December 2018.

Variable	Coefficient	SE	Z	p > |z|	95% LCL	95% UCL
Sex	–0.086	0.1017	–0.85	0.398	–0.285	0.113
Race/ethnicity	1.382	0.5551	2.49	0.013	0.294	2.470
Diagnosis information	–0.304	0.1250	–2.43	0.015	–0.549	–0.059
Age at diagnosis, years	0.077	0.0126	6.14	0.000	0.053	0.102
Symptom onset to diagnosis, years	–0.188	0.0345	–5.44	0.000	–0.255	–0.120
Race/ethnicity × age at diagnosis	–0.028	0.0090	–3.11	0.002	–0.046	–0.010

SE: standard deviation; LCL: lower limit confidence interval; UCL: upper limit confidence interval.

Significant p value <0.05.

Table 4 Estimated hazard ratios and 95% confidence intervals for the variable not involved with an interaction in the model from Table 2.

	Hazard ratio	95% CI
Sex
Male	1
Female	0.91	0.74, 1.12
Diagnosis information
Definite or probable	1.36	1.06, 1.73
Possible or unclassified	1
Symptom onset to diagnosis^a	0.83	0.77, 0.88

^a HR for a 12-month increase in the interval between symptom onset and diagnosis.

Discussion

Demographic analysis of Chicago and Detroit cases

As ALS is a non-notifiable disease in the U.S., there are limited survival data of ALS cases in a defined geographic cohort. This analysis combined detailed data collection of ALS cases reported by neurologists in a defined geographic area and followed them until death or the end of the analysis, 31 December 2018. This population-based cohort analysis has a high case-ascertainment rate with 99% of expected case reporting in Detroit and 88% of expected cases in Chicago. To our knowledge, this analysis is one of the first with a racially diverse group of ALS patients in the Chicago and Detroit cohort to provide incidence estimates and survival probability analysis followed through 31 December 2018.

The Chicago and Detroit metropolitan areas were selected and combined to overrepresent racial and ethnic minority populations due to their similar demographics. The counties from the Chicago and Detroit areas had a higher African American and Hispanic population compared with the U.S. (16,18). The demographic characteristics of ALS cases in the combined Chicago and Detroit metropolitan are similar to previously published literature with a larger percentage of cases being male, White, and diagnosed over the age of 60 (7,19). However, the percentage of ALS cases reported as African American is higher in the combined Chicago/Detroit cohort than the rates of cases in the United States (7).

Incidence analysis

Understanding estimates of incidence of ALS across the United States can help healthcare providers and agencies provide better care and services to those diagnosed with ALS. The average incidence averaged for the three-year study period was higher for Detroit (1.72 per 100,000) compared with Chicago with 1.54. This could be due to Chicago having a more diverse race and ethnicity demographic of cases compared with Detroit. The higher incidence rates for Detroit could also be driven by the higher incidence rate in the White population (2.41) compared with Chicago (1.50). The difference could also be attributed to capturing a slightly lower percentage of expected cases in Chicago compared with Detroit or missing/unknown race and ethnicity in both Chicago and Detroit. Furthermore, the Massachusetts Argeo Paul Cellucci ALS Registry, a state specific population-based ALS Registry, reports slightly higher incidence rates of 2.0–2.4 for 2009–2011. The higher incidence rates could be attributed to Massachusetts being the only state in the U.S. where ALS is a reportable disease and case ascertainment is derived directly from neurologists (20). Our incidence rates for Chicago/Detroit align with the overall pooled crude incidence rate of 1.75 per 100,000 PY derived from a global meta-analysis (14). A systematic review and meta-analysis from Xu et al. found an overall crude incidence rate of 1.59 per 100,000 PY (21). An additional study cites and reports a worldwide incidence rate of 2.08 per 100,000 PY with lower rates cited in the Hispanic, African, and Asian populations (8,22).

The combined Chicago and Detroit cohort had a higher percentage of non-White ALS cases compared with the national incidence rates. Our overall combined incidence rate of 1.56 per 100,000 PY is lower than the incidence rate of 2.6 per 100,000 PY across multiple European countries with ALS registries which has a less diverse population compared with other countries (23). Another study similarly found a crude incidence rate of 2.16 per 100,000 PY cross three European countries with a lesser diverse population (24). Similarly, those in the Chicago/Detroit combined cohort that reported their race as White had a significantly higher incidence rate compared with Asians or African Americans. A systematic review of ethnicity variation in ALS found lower incidence rates among African Americans, Asian, and Hispanic ethnicities compared with those who identified as White/Caucasian (25). A study from the Korean National Health Insurance Service (2011–2015) reported an overall incidence rate of 1.2 per 100,000 (26). A similar study conducted in Beijing from 2010 to 2015 estimated an annual incidence rate of 0.8/100,000 PY, which has similar results to studies conducted in Hong Kong and Taiwan (27). An additional review also suggests lower incidence rates in Asian populations compared with non-Asian populations (28). This aligns with the results of our analysis where the stratified Asian incidence rates for Chicago and Detroit are lower compared with the rates among the White and African American populations.

Furthermore, in a systematic review of ALS publications between 1966 and 2006, only 10% of studies assessed incidence rates stratified by ethnicity (8). This analysis stratified by race and ethnicity for the Chicago and Detroit populations adds to the current literature available about ALS cases. Additionally, the findings of this incidence and survival analysis adds to the current ALS literature and is crucial because of the projected probability of an increase in ALS cases from 2015 to 2040, due to an increase in the aging population (29).

Survival analysis

Survival estimates of this population were typical of other reports. Median survival for these participants was estimated to be 2.2 years, with more than 20% of participants surviving at least 10 years. Chio and colleagues report that median survival for PALS range from 20 to 48 months with 10–20% of PALS surviving longer than 10 years (6). While within reported ranges, there is the potential for this analysis to include participants experiencing slightly longer than typical survival, indicative of potential selection or measurement bias. Participants were selected from metropolitan areas where neurologists, medical centers, and medical schools are more prevalent, and thus more specialized care and earlier intervention are more likely than in rural areas. Patients going to a facility specializing in caring for PALS have longer survival than PALS from rural locations with less specialized care (30).

Similar to other studies, longer times between symptom onset and diagnosis predicted longer survival times (6). This is suspected to be an indication of slower progression and thus longer survival. Cases used in this analysis with definite or probable diagnosis experienced worsened survival; a diagnosis of definite ALS implies muscle weakness and wasting in more regions of the body relating to poorer prognosis (6,31–33). There were no statistically significant differences in survival among male and female PALS. Most studies have reported that male and female PALS do not experience differential survival, despite the higher reports of bulbar onset in female PALS with only a few exceptions (6,34). This requires further investigation to understand differences in survival between male and female PALS.

Comparable to previously published literature, this analysis reports worsening survival with increasing age at diagnosis (35). Of note, there is differential survival for PALS of different ages at diagnosis for White vs. non-White PALS, although the confidence bands for White and non-Whites greatly overlap for PALS under 70 years of age (Figure 2(a,b)). However, the example in Figure 2(c) shows generally improved survival among non-Whites at age 70. This is similar to estimates from Roberts et al. (34). Their examination of survival for White and non-White PALS reported longer survival for non-White PALS. What is interesting about the results of this analysis are the changes in survival depending on age of diagnosis. More investigations into the possible mechanisms behind this observation are needed. This could potentially be attributed to differences in income, lifestyle, or care provided by White and non-White families as the mechanism for differential survival. Ethnic minority familial caregivers, while having access to fewer resources, have been reported to provide more care and held stronger beliefs related to family obligation than their White counterparts (36). As with any surveillance and survival analysis, there are a few limitations for the findings of this analysis. An additional limitation to the findings of this analysis could be the number of cases in Chicago missing race and number of cases in Detroit missing ethnicity. Some of the data collection of the cases may be biased due to the larger proportion of referral centers participating in the case reporting.

Conclusion

With ALS being a non-notifiable disease in the U.S., conducting surveillance for this disease across the U.S. can be a challenge. The results of this analysis are consistent with other population-based cohort studies. However, this analysis was the first to use a racially diverse population in the Chicago and Detroit metropolitan area cohort to understand the incidence and survival characteristics of ALS patients. In addition, this analysis included one of the longer follow-up periods for survival time from 2009 through 2018. Knowing that race and increasing age at diagnosis have a more significant impact on survival can better guide neurologists and researchers. The results of this analysis add to the growing literature on survival of ALS cases across the U.S. Further investigation is needed on the impact of race on ALS survival in populations across the U.S. Better understanding the survival experience of ALS patients aids in understanding mutable prognostic factors, which can potentially extend survival and improve disease management.

Acknowledgements

The authors thank the state and local health departments and other organizations that assisted with data collection.

Declaration of interest

The authors declare no conflicts of interest. The findings and conclusions in this report are those of the author(s) and do not necessarily represent the official position of the Centers for Disease Control and Prevention/the Agency for Toxic Substances and Disease Registry and Prevention/the Agency for Toxic Substances and Disease Registry.

Additional information

Funding

This project was funded by the Agency for Toxic Substances and Disease Registry (ATSDR).