Hospital-treated infections in early- and mid-life and risk of Alzheimer's disease, Parkinson's disease, and amyotrophic lateral sclerosis: A nationwide nested case-control study in Sweden.

BACKGROUND: Experimental observations have suggested a role of infection in the etiology of neurodegenerative disease. In human studies, however, it is difficult to disentangle whether infection is a risk factor or rather a comorbidity or secondary event of neurodegenerative disease. To this end, we examined the risk of 3 most common neurodegenerative diseases in relation to previous inpatient or outpatient episodes of hospital-treated infections. METHODS AND
FINDINGS: We performed a nested case-control study based on several national registers in Sweden. Cases were individuals newly diagnosed with Alzheimer's disease (AD), Parkinson's disease (PD), or amyotrophic lateral sclerosis (ALS) during 1970 to 2016 in Sweden, identified from the National Patient Register. For each case, 5 controls individually matched to the case on sex and year of birth were randomly selected from the general population. Conditional logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) with adjustment for potential confounders, including sex, year of birth, area of residence, educational attainment, family history of neurodegenerative disease, and Charlson comorbidity index. Infections experienced within 5 years before diagnosis of neurodegenerative disease were excluded to reduce the influence of surveillance bias and reverse causation. The analysis included 291,941 AD cases (median age at diagnosis: 76.2 years; male: 46.6%), 103,919 PD cases (74.3; 55.1%), and 10,161 ALS cases (69.3; 56.8%). A hospital-treated infection 5 or more years earlier was associated with an increased risk of AD (OR = 1.16, 95% CI: 1.15 to 1.18, P < 0.001) and PD (OR = 1.04, 95% CI: 1.02 to 1.06, P < 0.001). Similar results were observed for bacterial, viral, and other infections and among different sites of infection including gastrointestinal and genitourinary infections. Multiple infections before age 40 conveyed the greatest risk of AD (OR = 2.62, 95% CI: 2.52 to 2.72, P < 0.001) and PD (OR = 1.41, 95% CI: 1.29 to 1.53, P < 0.001). The associations were primarily due to AD and PD diagnosed before 60 years (OR = 1.93, 95% CI: 1.89 to 1.98 for AD, P < 0.001; OR = 1.29, 95% CI: 1.22 to 1.36 for PD, P < 0.001), whereas no association was found for those diagnosed at 60 years or older (OR = 1.00, 95% CI: 0.98 to 1.01 for AD, P = 0.508; OR = 1.01, 95% CI: 0.99 to 1.03 for PD, P = 0.382). No association was observed for ALS (OR = 0.97, 95% CI: 0.92 to 1.03, P = 0.384), regardless of age at diagnosis. Excluding infections experienced within 10 years before diagnosis of neurodegenerative disease confirmed these findings. Study limitations include the potential misclassification of hospital-treated infections and neurodegenerative diseases due to incomplete coverage of the National Patient Register, as well as the residual confounding from unmeasured risk or protective factors for neurodegenerative diseases.
CONCLUSIONS: Hospital-treated infections, especially in early- and mid-life, were associated with an increased risk of AD and PD, primarily among AD and PD cases diagnosed before 60 years. These findings suggest that infectious events may be a trigger or amplifier of a preexisting disease process, leading to clinical onset of neurodegenerative disease at a relatively early age. However, due to the observational nature of the study, these results do not formally prove a causal link.

Introduction

Neurodegenerative diseases, including Alzheimer’s disease (AD), Parkinson’s disease (PD), and amyotrophic lateral sclerosis (ALS), are characterized by progressive loss of neurons in the nervous systems [1]. Although the incidence and prevalence varied with age, sex, and geography, the global burden of neurodegenerative diseases more than doubled during 1990 to 2016 [2-4]. Both genetic and nongenetic factors contribute to its development [5], but only a small proportion of patients with neurodegenerative diseases are driven by genetic causes [6,7].

A potential infectious etiology has been hypothesized for neurodegenerative diseases, as findings in animal studies have demonstrated that infectious processes might impact pathogenesis, phenotype, and progression of neurodegenerative disease [1,7-9]. Proposed underlying mechanisms include modulating misfolding and aggregation of pathological proteins, neuroinflammation, and infiltration of peripheral immune cells into the central nervous system (CNS) [1,9,10]. The extrapolation of such findings to a human context is, however, not straightforward. Previous studies have mostly examined the role of specific pathogens on a specific neurodegenerative disease, e.g., herpesvirus for AD [11], and influenza [12], hepatitis C virus [13], and Helicobacter pylori [14] for PD, with inconclusive results [7-9,11,15,16]. Although several studies have also assessed associations between infectious diseases and risk of dementia [17-19] and AD [20], influence of potential surveillance bias (greater-than-expected surveillance of disease after infections) and reverse causation (due to, for example, diagnostic delay of neurodegenerative diseases) on the associations was not always fully addressed. Therefore, whether infection is indeed a risk factor rather a comorbidity or secondary event of neurodegenerative disease remains unknown. In contrast to AD and PD, the potential link between infection and ALS has been less explored [8,21].

Weak evidence from human studies may be attributed to multiple factors. First, only few studies applied a prospective study design, making it difficult to differentiate causality from epiphenomena. Moreover, given that studies addressed different infections and neurodegenerative diseases, it is challenging to disentangle methodological drawbacks from real biological difference. Therefore, a comprehensive examination of different infections and different neurodegenerative diseases in a single study population might be of importance, after careful consideration of surveillance bias and reverse causation. Finally, a life course approach may also help to assess potential high-risk time window and critical periods for intervention [15].

To this end, using Swedish national healthcare registers, we performed a nationwide nested case-control study to examine the associations of hospital-treated infections, namely infections requiring inpatient or outpatient care, with the risk of AD, PD, and ALS. Our secondary aim was to explore whether the associations varied by type, site, age, and frequency of infection.

Methods

Study design

All individuals born after 1900 in Sweden whose parents were also born in Sweden were eligible for this study (N = 12,275,551). We followed these individuals from 1970 until a diagnosis of neurodegenerative disease, emigration, death, or December 31, 2016, whichever occurred first, through cross-linkages to the National Patient Register and the Causes of Death Register, using the individually unique Swedish personal identity number. The National Patient Register was established in 1964 and started to include all inpatient care since 1987; it also includes over 80% of outpatient care since 2001 [22]. We used this register to identify new diagnoses of neurodegenerative diseases via the Swedish revisions of the International Classification of Disease (ICD) codes, considering both primary and secondary diagnoses (S1 Table). Individuals with a diagnosis of neurodegenerative diseases before 1970 were excluded from the analyses, and individuals with multiple neurodegenerative diseases during follow-up contributed to the analyses of different diseases. The register-based definitions of AD, PD, and ALS have been validated against gold-standard clinical workup, showing a high specificity but a varying sensitivity and positive predictive value for AD (99.7%, 32.5%, and 57%) [23], PD (>98%, 72%, and 71%) [24], and ALS (all >90%) [25]. The relatively low sensitivity of AD diagnosis is likely attributable to misdiagnosis of AD as other dementias [23]. Date of diagnosis was defined as the date of first hospital visit concerning the disease as either the primary or a secondary diagnosis.

A nested case-control study within the above study base was then conducted to assess the associations of hospital-treated infections with the risk of AD, PD, and ALS. Five controls per case, individually matched by sex and year of birth, were randomly selected from the study base using the method of incidence density sampling [26]. Date of diagnosis and date of selection were used as the index date for cases and controls, respectively. Controls should be alive and free of the specific neurodegenerative disease of their matched case when selected. The prespecified analysis plan is presented in S1 Text.

Hospital-treated infections

Hospital inpatient and outpatient visits with a diagnosis of infection before index date were identified from the National Patient Register, using ICD codes shown in S2 Table [27]. We first studied any infection as a binary variable and then studied infections by type (bacterial, viral, or other infection), site (CNS, gastrointestinal, genitourinary, respiratory, or skin infection), age (<40, 40 to 59.9, or ≥60 years), and frequency (0, 1, or ≥2 events).

Covariates

We identified information on area of residence (3 groups: Northern, Central, and Southern Sweden) from the Total Population Register (information available from 1947 onward) and educational attainment (4 groups: 0 to 9 years, 10 to 12 years, ≥13 years, and “missing”) from the Swedish Longitudinal Integrated Database for Health Insurance and Labour Market Studies (information available from 1990 onward) [28]. Family history of neurodegenerative disease (yes/no) was defined as a diagnosis of the disease among first-degree relatives (i.e., biological parents and full siblings) (information available from 1964 onward). We identified parents and full siblings from the Swedish Multi-Generation Register [29]. Charlson comorbidity index (3 groups: 0, 1, and ≥2), as a proxy of general health status, was calculated using data from the National Patient Register (information available from 1964 onward) [30]. All covariates were measured on the index date of the cases and controls.

Statistical analyses

We performed separate analyses for AD, PD, and ALS. If 1 individual was diagnosed with more than one of the studied neurodegenerative diseases, for example, a diagnosis of PD first and a diagnosis of AD later, this individual would contribute to the analysis of PD and AD with the respective diagnosis date. We first described the percentage of individuals with a history of infections among cases and controls during 20 years before index date, and then applied conditional logistic regression to estimate odds ratio (OR) and 95% confidence interval (CI), as an estimate of the association between infections and risk of neurodegenerative disease. As OR obtained from a nested case-control study within a well-defined underlying cohort and with controls selected using the method of incidence density sampling is mathematically equal to relative risk estimate of the underlying cohort study [31]. OR in the present study can be interpreted as incidence rate ratio [32]. In addition to conditioning on matching variables (sex and year of birth), we adjusted for area of residence, educational attainment, family history of neurodegenerative disease, and Charlson comorbidity index (i.e., history of comorbidity) in the analysis. Due to the concern of surveillance bias and reverse causation [23,24,33], we used a lag time of 5 years, namely all infections experienced during 5 years before the index date were excluded from the analysis.

The analyses were first performed for any infection and then by type, site, age, and frequency of infection. To examine the potential dose–response relationship within specific age groups, we also analyzed frequency of infection by age at infection. We stratified the analysis by sex (male or female), age at index date (<60 years or ≥60 years), calendar period (1970–1986, 1987–2000, and 2001–2016), and year of birth (1900 to 1919, 1920 to 1939, 1940 to 1959, or ≥1960) to assess whether the associations would differ by sex, age, calendar period, and birth cohort. As previous studies have used age 65 to define early- and late-onset AD [34], we additionally performed a stratified analysis of AD by age 65 at index date. The P value for interaction was calculated using the Wald test for the product terms between infection and a specific covariate. Because the positive associations were mainly noted for those diagnosed before 60 years, and the etiologies for AD and PD diagnosed at relatively young age are potentially different from those diagnosed at later age, we repeated all analyses for AD, PD, and ALS diagnosed before or after 60 years, separately.

We also performed a series of secondary analyses to assess the robustness of the findings. (1) To assess whether the associations could be modified by family history of neurodegenerative disease, in a sensitivity analysis, we restricted the analysis to those without a family history of the disease. (2) To minimize potential misclassification of outcomes, we, in another sensitivity analysis, restricted the definition of outcomes to those with at least 2 hospital visits concerning the same disease. (3) Because information on education was available only from 1990 onward, there was a high degree of missingness for this covariate. To assess to what extent this could influence the results, we repeated the main analysis without adjustment for education and additionally compared the results between 2 models with or without adjustment for education among individuals with data available on education. (4) We additionally excluded individuals with more than 1 neurodegenerative disease from the analysis (2.64% with AD and PD, 0.05% with AD and ALS, 0.04% with PD and ALS, and <0.01% with all 3 diseases). (5) Finally, to assess the impact of choice of lag time, we performed another sensitivity analysis by excluding infections experienced during 10 years before the index date.

Data analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC) and R version 3.6.0. A two-sided P ≤ 0.05 was considered statistically significant. This study is reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline (S1 STROBE Checklist).

Ethics consideration

The study was approved by the Regional Ethical Review Board in Stockholm (2012/1814-31/4). According to the current Swedish regulation, informed consent is not required when register data are used for the purpose of research.

Results

We enrolled 291,941 AD cases (median age at diagnosis: 76.2 years; male: 46.6%), 103,919 PD cases (74.3; 55.1%), and 10,161 ALS cases (69.3; 56.8%), together with their matched controls in the study (Table 1). Compared with controls, individuals with neurodegenerative diseases had a higher percentage of family history of neurodegenerative disease but lower Charlson comorbidity index score.

Table 1

Characteristics of patients with neurodegenerative diseases and their matched controls at the index date.

	AD		PD		ALS
	Case	Control	Case	Control	Case	Control
N	291,941	1,459,705	103,919	519,595	10,161	50,805
Age at the index date, years
Mean ± SD	67.7 ± 21.2	67.7 ± 21.2	71.5 ± 13.8	71.5 ± 13.8	67.7 ± 12.1	67.7 ± 12.1
Median (Q1-Q3)	76.2 (58.0–82.6)	76.2 (58.0–82.6)	74.3 (67.1–79.9)	74.3 (67.1–79.9)	69.3 (61.3–76.1)	69.3 (61.4–76.1)
<60 years, n (%)	76,023 (26.0)	380,101 (26.0)	13,319 (12.8)	66,661 (12.8)	2,245 (22.1)	11,216 (22.1)
≥60 years, n (%)	215,918 (74.0)	1,079,604 (74.0)	90,600 (87.2)	452,934 (87.2)	7,916 (77.9)	39,589 (77.9)
Sex, n (%)
Male	135,998 (46.6)	679,990 (46.6)	57,281 (55.1)	286,405 (55.1)	5,773 (56.8)	28,865 (56.8)
Female	155,943 (53.4)	779,715 (53.4)	46,638 (44.9)	233,190 (44.9)	4,388 (43.2)	21,940 (43.2)
Area of residence, n (%)
Northern Sweden	65,230 (22.3)	353,015 (24.2)	25,162 (24.2)	125,333 (24.1)	2,316 (22.8)	12,224 (24.1)
Central Sweden	162,484 (55.7)	775,595 (53.1)	55,654 (53.6)	275,688 (53.1)	5,589 (55.0)	26,994 (53.1)
Southern Sweden	64,227 (22.0)	331,095 (22.7)	23,103 (22.2)	118,574 (22.8)	2,256 (22.2)	11,587 (22.8)
Educational attainment, n (%)
0–9 years	77,523 (26.6)	388,391 (26.6)	29,081 (28.0)	153,830 (29.6)	2,922 (28.8)	15,586 (30.7)
10–12 years	45,667 (15.6)	219,121 (15.0)	18,903 (18.2)	92,500 (17.8)	2,725 (26.8)	12,694 (25.0)
≥13 years	16,711 (5.7)	91,564 (6.3)	10,076 (9.7)	42,998 (8.3)	1,393 (13.7)	6,751 (13.3)
Missing	152,040 (52.1)	760,629 (52.1)	45,859 (44.1)	230,267 (44.3)	3,121 (30.7)	15,774 (31.1)
Family history of the disease, n (%)	18,345 (6.3)	54,522 (3.7)	2,377 (2.3)	7,088 (1.4)	181 (1.8)	126 (0.3)
Charlson comorbidity index, n (%)
0	251,119 (86.0)	1,220,639 (83.6)	87,590 (84.3)	436,444 (84.0)	8,834 (86.9)	43,972 (86.6)
1	37,134 (12.7)	215,219 (14.7)	14,604 (14.1)	74,907 (14.4)	1,197 (11.8)	6,171 (12.2)
≥2	3,688 (1.3)	23,847 (1.6)	1,725 (1.7)	8,244 (1.6)	130 (1.3)	662 (1.3)

AD, Alzheimer’s disease; ALS, amyotrophic lateral sclerosis; PD, Parkinson’s disease; SD, standard deviation.

Index date: date of diagnosis for cases and date of selection for controls.

Compared with controls, there was a slightly higher percentage of individuals with hospital-treated infections among patients with AD or PD, but not ALS, during the 20 years before the index date (Fig 1). The percentage increased rapidly during the year before diagnosis of neurodegenerative disease.

Fig 1

Percentage of individuals with hospital-treated infections among patients with neurodegenerative diseases and their matched controls during the 20 years before the index date.

AD, Alzheimer’s disease; ALS, amyotrophic lateral sclerosis; PD, Parkinson’s disease.

Primary analyses

After excluding infections diagnosed during 5 years before the index date, an event of hospital-treated infection was associated with a higher risk of AD (OR = 1.16, 95% CI: 1.15 to 1.18, P < 0.001) and PD (OR = 1.04, 95% CI: 1.02 to 1.06, P < 0.001), but not ALS (OR = 0.97, 95% CI: 0.92 to 1.03, P = 0.384) (Table 2). Positive associations for AD and PD were similarly observed for bacterial, viral, and other infections, as well as for gastrointestinal and genitourinary infections (Table 2). For ALS, we did not observe any association.

Table 2

Association between hospital-treated infection and risk of neurodegenerative disease.

	AD				PD				ALS
Group	Infection (case/control)	No infection (case/control)	OR (95% CI)	P	Infection (case/control)	No infection (case/control)	OR (95% CI)	P	Infection (case/control)	No infection (case/control)	OR (95% CI)	P
Any infection	49,789/225,776	242,152/1,233,929	1.16 (1.15–1.18)	<0.001	16,818/81,891	87,101/437,704	1.04 (1.02–1.06)	<0.001	1,793/9,184	8,368/41,621	0.97 (0.92–1.03)	0.384
Infection type
Bacterial	27,182/120,724	264,759/1,338,981	1.16 (1.15–1.18)	<0.001	9,113/44,038	94,806/475,557	1.04 (1.02–1.07)	0.001	983/4949	9,178/45,856	1.00 (0.93–1.08)	0.970
Viral	14,136/64,001	277,805/1,395,704	1.12 (1.10–1.15)	<0.001	4,860/22,910	99,059/496,685	1.07 (1.03–1.10)	<0.001	549/2,618	9,612/48,187	1.05 (0.95–1.16)	0.324
Others	5,336/23,658	286,605/1,436,047	1.15 (1.11–1.18)	<0.001	1,886/8,826	102,033/510,769	1.07 (1.02–1.13)	0.006	252/1,165	9,909/49,640	1.09 (0.95–1.26)	0.219
Infection site
CNS	1,865/8,354	290,076/1,451,351	1.13 (1.07–1.19)	<0.001	672/3,133	103,247/516,462	1.07 (0.99–1.17)	0.097	78/436	10,083/50,369	0.89 (0.69–1.13)	0.335
Gastrointestinal	9,038/39,856	282,903/1,419,849	1.23 (1.20–1.26)	<0.001	3,041/14,386	100,878/505,209	1.07 (1.02–1.11)	0.002	326/1,776	9,835/49,029	0.92 (0.81–1.04)	0.188
Respiratory	17,649/85,976	274,292/1,373,729	1.04 (1.03–1.06)	<0.001	6,087/30,823	97,832/488,772	0.99 (0.96–1.02)	0.485	666/3,153	9,495/47,652	1.07 (0.98–1.17)	0.145
Genitourinary	5,754/25,730	286,187/1,433,975	1.15 (1.11–1.18)	<0.001	1,891/8,258	102,028/511,337	1.15 (1.09–1.21)	<0.001	163/807	9,998/49,998	1.02 (0.86–1.21)	0.794
Skin	4,345/19,069	287,596/1,440,636	1.15 (1.12–1.19)	<0.001	1,450/7,232	102,469/512,363	1.01 (0.95–1.07)	0.813	181/847	9,980/49,958	1.06 (0.90–1.25)	0.480

AD, Alzheimer’s disease; ALS, amyotrophic lateral sclerosis; CI, confidence interval; CNS, the central nervous system; OR, odds ratio; PD, Parkinson’s disease.

Conditional on matching factors (sex and year of birth) and further adjusted for area of residence, educational attainment, family history of the disease, and history of comorbidity. Infections diagnosed during 5 years before the index date were excluded to alleviate the potential influence of reverse causation due to diagnostic delay and surveillance bias.

Magnitude of the associations of any hospital-treated infections with the risk for AD and PD decreased with increasing age at infection, with the strongest associations noted for infection at age 40 or below (OR = 1.86, 95% CI: 1.82 to 1.90 for AD, P < 0.001; OR = 1.20, 95% CI: 1.15 to 1.25 for PD, P < 0.001) (S3 Table). Dose–response relationships were observed for frequency of infections before 40 years (all P < 0.001). For example, individuals with ≥2 events of infections before 40 years had the highest risk of AD (OR = 2.62, 95% CI: 2.52 to 2.72, P < 0.001) and PD (OR = 1.41, 95% CI: 1.29 to 1.53, P < 0.001) (S3 Table).

In the stratified analyses of any hospital-treated infection, a stronger association was observed among male (OR = 1.20, 95% CI: 1.18 to 1.22, P < 0.001) than female (OR = 1.13, 95% CI: 1.12 to 1.15, P < 0.001) for AD, but among female (OR = 1.07, 95% CI: 1.05 to 1.11, P < 0.001) than male (OR = 1.01, 95% CI: 0.98 to 1.03, P = 0.536) for PD (both P < 0.001, S4 Table). The associations were primarily limited to individuals diagnosed before 60 years (OR = 1.93, 95% CI: 1.89 to 1.98 for AD, P < 0.001; OR = 1.29, 95% CI: 1.22 to 1.36 for PD, P < 0.001), and were stronger during 1970 to 1986 (OR = 1.32, 95% CI: 1.27 to 1.36 for AD, P < 0.001; OR = 1.14, 95% CI: 1.08 to 1.20 for PD, P < 0.001), and among individuals born after 1960 (OR = 1.59, 95% CI: 1.52 to 1.65 for AD, P < 0.001; OR = 1.39, 95% CI: 1.24 to 1.56 for PD, P < 0.001) (S4 Table). Stronger associations were also observed for AD ascertained before 65 (e.g., for individuals with infections before 40 years, OR = 1.95, 95% CI: 1.91 to 2.00, P < 0.001), compared with AD ascertained after 65 (OR = 1.11, 95% CI: 1.05 to 1.18, P < 0.001) (S5 Table).

All these results were primarily attributable to AD and PD diagnosed before 60 years (Table 3), compared with those diagnosed at 60 years or older (S6 Table). For example, genitourinary infection was associated with the highest risk of AD before 60 years (OR = 2.75, 95% CI: 2.54 to 2.98, P < 0.001), whereas CNS infection was associated the highest risk of PD before 60 years (OR = 1.39, 95% CI: 1.15 to 1.69, P < 0.001). Gastrointestinal infection was also associated with a higher risk of AD (OR = 1.35, 95% CI: 1.28 to 1.41, P < 0.001) and PD (OR = 1.18, 95% CI: 1.06 to 1.32, P < 0.001) diagnosed before 60 years.

Table 3

Association between hospital-treated infection and risk of neurodegenerative disease diagnosed before 60 years.

Group	AD		PD		ALS
	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P
Analyses by characteristics of infections
Infection type
Bacterial	2.08 (2.02–2.15)	<0.001	1.30 (1.21–1.39)	<0.001	1.09 (0.93–1.28)	0.275
Viral	1.87 (1.79–1.94)	<0.001	1.36 (1.24–1.49)	<0.001	1.02 (0.84–1.24)	0.816
Others	1.85 (1.73–1.98)	<0.001	1.33 (1.16–1.53)	<0.001	0.97 (0.71–1.33)	0.844
Infection site
CNS	1.63 (1.49–1.78)	<0.001	1.39 (1.15–1.69)	0.001	1.13 (0.78–1.64)	0.516
Gastrointestinal	1.35 (1.28–1.41)	<0.001	1.18 (1.06–1.32)	0.002	0.92 (0.72–1.17)	0.494
Respiratory	1.63 (1.57–1.70)	<0.001	1.33 (1.22–1.46)	<0.001	1.12 (0.93–1.35)	0.217
Genitourinary	2.75 (2.54–2.98)	<0.001	1.26 (1.04–1.52)	0.016	1.29 (0.86–1.92)	0.213
Skin	1.89 (1.77–2.02)	<0.001	1.16 (0.99–1.36)	0.071	1.50 (1.12–2.02)	0.007
Age at infection
<40 years	1.95 (1.90–2.00)	<0.001	1.36 (1.28–1.45)	<0.001	0.98 (0.84–1.13)	0.731
40–59.9 years	1.72 (1.62–1.82)	<0.001	1.08 (0.99–1.18)	0.093	1.03 (0.84–1.25)	0.797
Age and frequency of infection
<40 years
0	Ref.		Ref.		Ref.
1	1.73 (1.69–1.78)	<0.001	1.27 (1.18–1.37)	<0.001	0.96 (0.81–1.13)	0.595
≥2	2.70 (2.59–2.82)	<0.001	1.63 (1.47–1.82)	<0.001	1.02 (0.81–1.28)	0.873
40–59.9 years
0	Ref.		Ref.		Ref.
1	1.64 (1.53–1.76)	<0.001	1.02 (0.92–1.14)	0.692	1.13 (0.90–1.42)	0.280
≥ 2	1.97 (1.76–2.20)	<0.001	1.21 (1.05–1.41)	0.011	0.80 (0.55–1.15)	0.229
Stratified analyses by sex, calendar period, and birth cohort
Sex
Male	1.76 (1.70–1.81)	<0.001	1.17 (1.08–1.27)	<0.001	1.07 (0.90–1.26)	0.468
Female	2.18 (2.11–2.26)	<0.001	1.42 (1.31–1.54)	<0.001	0.98 (0.81–1.19)	0.839
Calendar period at diagnosis
1970–1986	2.39 (2.26–2.52)	<0.001	1.42 (1.25–1.62)	<0.001	1.22 (0.80–1.86)	0.360
1987–2000	1.88 (1.83–1.93)	<0.001	1.59 (1.45–1.75)	<0.001	0.80 (0.59–1.10)	0.170
2001–2016	1.35 (1.20–1.52)	<0.001	1.04 (0.95–1.13)	0.386	1.06 (0.91–1.23)	0.458
Year of birth
1900–1919	2.08 (1.35–3.19)	0.001	2.59 (1.49–4.48)	0.001	2.30 (0.69–7.68)	0.174
1920–1939	1.87 (1.75–1.99)	<0.001	1.30 (1.14–1.48)	<0.001	0.88 (0.60–1.31)	0.539
1940–1959	2.15 (2.08–2.23)	<0.001	1.14 (1.05–1.23)	0.001	0.97 (0.81–1.15)	0.694
≥ 1960	1.59 (1.52–1.65)	<0.001	1.39 (1.24–1.56)	<0.001	1.11 (0.88–1.40)	0.366

AD, Alzheimer’s disease; ALS, amyotrophic lateral sclerosis; CI, confidence interval; CNS, the central nervous system; OR, odds ratio; PD, Parkinson’s disease.

Conditional on matching factors (age and sex) and further adjusted for area of residence, educational attainment, family history of the disease, and history of comorbidity. Infections diagnosed during 5 years before the index date were excluded to alleviate the potential influence of reverse causation due to diagnostic delay.

Secondary analyses

Compared to the main analyses, similar results were observed when restricting the analyses to individuals without a family history of the disease or when defining neurodegenerative diseases through at least 2 hospital visits concerning the same disease (S7 Table). We observed similar results in the model without adjusting for education and when restricting the analysis to individuals with complete data on education, consistent results were also noted between the models with or without adjustment for education (S7 Table). Similar results were also observed after excluding individuals with multiple neurodegenerative diseases (S7 Table). Finally, after excluding infections experienced during 10 years before index date, similar associations were noted between any and specific hospital-treated infections and risk of AD and PD (S8 Table). For example, for any infection, the OR of AD and PD was 1.18 (95% CI: 1.16 to 1.19, P < 0.001) and 1.04 (95% CI: 1.01 to 1.06, P = 0.002), respectively. A dose–response relationship was still observed for frequency of infections after using a 10-year lag time (P < 0.001). The OR of AD and PD was 2.22 (95% CI: 2.12 to 2.33, P < 0.001) and 1.26 (95% CI: 1.15 to 1.38, P < 0.001) in relation to ≥2 events of infection before age 40 (S9 Table).

Discussion

In this nationwide study, we found that hospital-treated infections 5 years or more prior to neurodegenerative disease diagnosis were associated with an increased risk of AD and PD, specifically among cases diagnosed before 60 years, but not ALS. The associations for AD and PD were observed across infection types and sites but were stronger for infections—especially repeated infections—in early- or mid-life.

Though causality cannot be inferred from the study, the similar results for different infection types and sites might suggest that the potential underlying mechanisms of the observed associations are not specific to certain pathogens and raise the possibility that systemic inflammation might play a role in brain health, an idea that is supported by evidence from a previous study reporting a relationship between hospital-treated infection and vascular dementia and AD [17]. Although unexpected, our subsequent findings that hospital-treated infections were more strongly associated with risk of AD and PD before 60 years, compared to later, and that individuals with repeated infections in early- and mid-life had the greatest risk increment of AD and PD, are new and potentially important. We hypothesize that infectious events may be a trigger or amplifier of a preexisting disease process, leading to clinical onset of neurodegenerative disease at a relatively early age among individuals with disease predisposition [35-37]. Active monitoring and prevention of severe infections may therefore help to prevent or delay disease onset among high-risk individuals.

The association between infections and an increased risk of AD has support from previous studies. Evidence has suggested a link between herpes virus infection, including herpes simplex virus type 1 and varicella zoster virus, and AD and dementia [11], although a review showed that available findings were inconsistent with overall low-quality evidence [16]. Cognitive impairment has also been shown among individuals with neurological varicella zoster virus infection or Coronavirus Disease 2019 (COVID-19) [38,39]. However, a reduced long-term risk of dementia was observed among individuals with herpes infection treated with antiviral medications [40]. Another study reported a dose–response relationship between number of infectious diseases and dementia risk [17]. This study, however, did not explore the effect of age at infection and whether a dose–response relationship existed still for AD after considering the influence of potential reverse causation due to the preclinical stage and diagnostic delay of AD. Two studies used the same datasets from the UK reported an increased risk of dementia [18] or AD [20] among individuals with infectious diseases, without, however, reporting a clear dose–response relationship.

Similarly, a link between infections and PD has also been proposed. Previous studies have suggested a role of influenza [12] and hepatitis C virus [13] on the risk of PD, whereas antiviral treatment of hepatitis C virus infection might mitigate such risk [41]. Occurrence of parkinsonism after influenza and worsening of parkinsonian syndromes following COVID-19 have also been described [42]. H. pylori infection was shown to be associated with an increased risk of PD [14,43], whereas eradication of this bacteria was shown to improve PD symptoms [44,45]. A recent study from Denmark reported an increased risk of PD with similar magnitude as the present study more than 10 years after exposure to infections; however, the result was not statistically significant (OR = 1.04, 95% CI: 0.98 to 1.10) [12]. These findings, together with the present study, support potential involvement of infection in the etiopathogenesis of AD and PD, although efforts are still needed to explain contradicting findings between studies [7,11,16].

In the present study, although positive associations were observed for AD and PD diagnosed before 60 years and for infections in early- and mid-life throughout the study period, the magnitude of the associations decreased over calendar period. There are likely multiple explanations. First, the treatment of infections might have changed during the study period, with presumably better treatment outcomes during later calendar periods than before. Similarly, the diagnosis of neurodegenerative diseases might have improved over calendar period, not only thanks to improvements in diagnostic tools, especially imaging and more recently also soluble biomarkers, but also due to improved societal awareness of these diseases and resources in neurology and geriatric care. Second, we included outpatient care data for the definition of hospital-treated infections since 2001, whereas only inpatient care data were available before then. Infections requiring specialized care, especially inpatient care, might have a different role on neurodegenerative diseases compared to infections not requiring such. Indeed, a previous study suggested a much weaker association between primary care–based infections, compared with hospital-treated infections, and the risk of dementia (hazard ratio = 1.02; 95% CI: 1.00 to 1.04 versus 1.99; 95% CI: 1.94 to 2.04) [18].

Our study did not support an association of hospital-treated infections with ALS risk. The null finding, however, does not rule out the possibility that milder infections not attended by specialist care might still be of importance. Previous studies have, for example, shown that infections might contribute to protein aggregation and mislocalization as well as glutamate excitotoxicity—known pathological processes of ALS [9]. Enterovirus RNA sequences have been found in the CNS of ALS patients [21,46], while disturbed gut microbiome composition [8] and increased use of antibiotics have also been observed among ALS patients [47].

Several mechanisms might explain the link between infection and neurodegenerative disease. As shown in animal research, infectious agents and their metabolites can promote aggregation of misfolded protein in neuron and its propagation from the periphery to the CNS, e.g., amyloid precursor protein and hyperphosphorylated tau protein in AD and alpha-synuclein in PD [1]. Infectious agents might also elicit inflammatory responses at infection site, resulting in production of pro-inflammatory cytokines and chemokines, which, like infectious agents, can cross the blood–brain barrier, enter the CNS, and elicit neuroinflammation through activating microglia and astrocytes [9]. Some neurotropic microbes, especially those with ineradicable infection, like herpesvirus, could not only trigger chronic neuroinflammation [48] but also directly infect neurons [7]. Prodromal nonmotor symptoms, including gastrointestinal symptoms and poor olfaction, have indeed been reported in PD patients years before onset of motor symptoms [49], whereas a lower PD risk has been suggested after vagotomy [50] and appendectomy [51].

Strengths and limitations

Unlike most previous studies, our study did not focus on specific infectious diseases (e.g., influenza, pneumonia, or viral hepatitis), but instead studied all infections requiring hospital treatment. To our knowledge, it is the first to date to assess the associations of hospital-treated infections—by type, site, age, and frequency—with risks of the 3 most common neurodegenerative diseases in the same population. Strengths of the study include the nationwide study design, large sample size, representativeness of patients with neurodegenerative diseases, and individually matched controls randomly selected from the general population. These strengths enabled us to perform informative subgroup analyses to demonstrate disease-specific results. Complete follow-up due to linkages to the national registers and the objective and prospective ascertainment of infection and neurodegenerative diseases minimized further selection and measurement biases commonly existent in observational studies.

There are also limitations. An infection might be secondary to undiagnosed neurodegenerative disease, especially during the years before diagnosis (reverse causation). Moreover, individuals with an ongoing infection might have a higher probability of being investigated for neurodegenerative disease, compared with others (surveillance bias). To address these concerns, we excluded infections experienced during 5 years before diagnosis in the main analysis and those experienced during 10 years before diagnosis in a sensitivity analysis, which rendered similar results.

Misclassification of hospital-treated infections and neurodegenerative diseases is another concern. Due to incomplete coverage of inpatient care data before 1987 and lack of outpatient care data before 2001 in the National Patient Register, some individuals with hospital-treated infections might have been misclassified as not having infection. Further, our definition of infections did not include those attended by primary care or not attended by healthcare at all. As a result, the present findings should only be interpreted in the context of relatively severe infections requiring specialist care. Besides, although register-based definition of infections was shown to have a high specificity (>95%) [22], we had no laboratory data to confirm the infections, due to the register-based nature of the study. In addition, as the sensitivity of register-based definition is low for AD (32.5%) [23], although higher for PD (72%) [24] and ALS (>90%) [25], not all patients with neurodegenerative diseases were identified, which might have diluted the real associations toward null.

The patient population included in the study consisted of all patients with a newly diagnosed neurodegenerative disease during 1970 to 2016 in Sweden. As suggested in previous validation studies [23-25], the National Patient Register is a valid data source for epidemiological studies of neurodegenerative diseases. However, as all cases were identified through specialist care, there was likely a delay comparing the register-based date of diagnosis (especially for the ones identified through inpatient care) and the actual date of diagnosis. Also, although the studied neurodegenerative diseases should theoretically all be referred to specialized care by a neurologist or geriatrician, we cannot exclude the possibility that some patients might have only been attended in primary care and misclassified as not having the disease in the National Patient Register. In addition, as we had relatively young individuals in the study, there was a high proportion of neurodegenerative diseases ascertained before 60 years. For example, 26% of the AD individuals were ascertained before 60 years (median age at diagnosis: 41; 56.4% of them born after 1956), higher than expected. However, the proportion was 13.3% among individuals born before 1956.

Residual confounding is also a concern as we did not have complete information on all risk or protective factors for neurodegenerative diseases, including lifestyle factors (smoking and body mass index), medical factors (e.g., human leucocyte antigen type [52] and brain damage due to vascular, traumatic, and illicit drug use reasons [53]), and genetic factors (e.g., ethnicity or risk genes [54]). Moreover, due to varying incidence of infections and neurodegenerative diseases across countries [2,3,8] and relatively young participants in present study, generalization of our findings to other settings should be done with caution. Similarly, whether hospital-treated infections indeed do not influence the risk of AD and PD beyond 60 years of age needs to be studied further. Finally, future studies are needed to better understand roles of specific pathogens, infectious duration, and treatments (e.g., antibiotics, known to affect microbial environment and lead to dysbiosis [7]) on the link between infections and neurodegenerative disease.

In conclusion, our study suggested that individuals with hospital-treated infections, especially in those occurring in early- and mid-life, had an increased risk of developing AD and PD, attributable to cases diagnosed before 60 years. Further studies are warranted to validate these findings, to elucidate underlying mechanisms, and to determine whether better control of hospital-treated infections could prevent or delay onset of neurodegenerative diseases, especially the ones with an onset relatively early in life.

STROBE statement—checklist of items that should be included in reports of observational studies.

(DOC)

Click here for additional data file.

The Swedish revisions of International Classification of Diseases (ICD) codes for neurodegenerative diseases.

(DOCX)

Click here for additional data file.

The Swedish revisions of International Classification of Diseases (ICD) codes for hospital-treated infections.

(DOCX)

Click here for additional data file.

Associations between age at hospital-treated infection and the consequent risks of neurodegenerative diseases (5-year lag time).

(DOCX)

Click here for additional data file.

Association between any hospital-treated infection and risk of neurodegenerative disease, stratified analyses by sex, age at diagnosis, calendar period at diagnosis, and year of birth (5-year lag time).

(DOCX)

Click here for additional data file.

Associations between age at hospital-treated infection and the consequent risks of early-onset AD and late-onset AD (5-year lag time).

(DOCX)

Click here for additional data file.

Association between hospital-treated infection and risk of neurodegenerative disease diagnosed at 60 years or older.

(DOCX)

Click here for additional data file.

Sensitivity analyses for the association between any hospital-treated infection and risk of neurodegenerative disease (5-year lag time).

(DOCX)

Click here for additional data file.

Association between any hospital-treated infection and risk of neurodegenerative disease (10-year lag time).

(DOCX)

Click here for additional data file.

Associations between age at hospital-treated infection and the consequent risks of neurodegenerative diseases (10-year lag time).

(DOCX)

Click here for additional data file.

Analysis plan.

(DOCX)

Click here for additional data file.

7 Apr 2022

Dear Dr Sun,

Thank you for submitting your manuscript entitled "Hospital-treated infections in early and mid-life increase the risk of Alzheimer’s disease and Parkinson’s disease" for consideration by PLOS Medicine.

Your manuscript has now been evaluated by the PLOS Medicine editorial staff and I am writing to let you know that we would like to send your submission out for external assessment.

However, we first need you to complete your submission by providing the metadata that is required. To this end, please login to Editorial Manager where you will find the paper in the 'Submissions Needing Revisions' folder on your homepage. Please click 'Revise Submission' from the Action Links and complete all additional questions in the submission questionnaire.

Please re-submit your manuscript within two working days, i.e. by Apr 11 2022 11:59PM.

Login to Editorial Manager here: https://www.editorialmanager.com/pmedicine

Once your full submission is complete, your paper will undergo a series of checks in preparation for external assessment.

Feel free to email us at plosmedicine@plos.org if you have any queries relating to your submission.

Kind regards,

Richard Turner, PhD

Senior Editor, PLOS Medicine

rturner@plos.org

8 May 2022

Dear Dr. Sun,

Thank you very much for submitting your manuscript "Hospital-treated infections in early and mid-life increase the risk of Alzheimer’s disease and Parkinson’s disease" (PMEDICINE-D-22-01054R1) for consideration at PLOS Medicine.

Your paper was evaluated by an academic editor with relevant expertise and sent to independent reviewers, including a statistical reviewer. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

In light of these reviews, we will not be able to accept the manuscript for publication in the journal in its current form, but we would like to invite you to submit a revised version that addresses the reviewers' and editors' comments fully. You will appreciate that we cannot make a decision about publication until we have seen the revised manuscript and your response, and we expect to seek re-review by one or more of the reviewers.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We hope to receive your revised manuscript by May 30 2022 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

Please let me know if you have any questions, and we look forward to receiving your revised manuscript.

Sincerely,

Richard Turner, PhD

Senior editor, PLOS Medicine

rturner@plos.org

-----------------------------------------------------------

Requests from the editors:

Please revisit the data statement (submission form) to detail the source(s) of data and provide contact information for readers interested in seeking access.

Please adapt the title to better match journal style: we suggest "Hospital-treated infections in early- and mid-life and risk of Alzheimer’s disease, Parkinson’s disease and amyotrophic lateral sclerosis: A case-control study".

Please specify the country in the title.

Please restructure the abstract to combine the "Methods" and "Findings" subsections.

Please add a new final sentence to the combined subsection, which should begin "Study limitations include ..." or similar and should quote 2-3 of the study's main limitations.

In the abstract and throughout the paper, please quote p values alongside 95% CI, where available.

After the abstract, please include and new and accessible "author summary" section in non-identical prose. You may find it helpful to consult one or two recent research articles published in PLOS Medicine to get a sense of the preferred style.

Early in the Methods section (main text), please state whether or not the study had a protocol or prespecified analysis plan, and if so attach the relevant document as a supplementary file.

Please highlight analyses that were not prespecified.

Throughout the text, please format reference call-outs to the following style: "... genetic causes [6,7]." (noting the absence of spaces within the square brackets).

In the reference list, please convert all italics and boldface text into plain text.

Where appropriate, please list 6 author names, followed by "et al.".

Noting reference 5 and others, please ensure that all references have full access details.

Please include a completed checklist for the most appropriate reporting guideline, e.g., STROBE, as an attachment, labelled "S1_STROBE_Checklist" or similar and referred to as such in the Methods section (main text).

In the checklist, please refer to individual items by section, e.g., "Methods" and paragraph number, not by line or page numbers as these generally change in the event of publication.

Comments from the reviewers:

*** Reviewer #1 (statistical reviewer):

This study aims to examine the risk of three most common neurodegenerative diseases in relation to previous inpatient or outpatient episodes of hospital-treated infections.

Comments:

"Hospital-treated infections, especially in early and mid-life, were associated with an increased subsequent risk of AD and PD, especially of AD and among cases diagnosed at relatively young age. These findings suggest that infectious events may be a trigger or amplifier of a pre existing disease process, leading to clinical onset of neurodegenerative disease at a relatively early age."

Can the authors please revisit the conclusions, being mindful that the study and analytical design does not allow causality to be inferred?

"All individuals born after 1900 in Sweden whose parents were also born in Sweden were eligible for this study (N=12,275,551). We followed them from 1970 until a diagnosis of neurodegenerative disease, emigration, death, or December 31st, 2016, whichever occurred first, through cross linkages to the National Patient Register (NPR) and the Causes of Death Register, using the unique personal identity number."

Can the authors please clarify if they did or did not capture cases of onset of neurodegenerative disease before 1970?

Can the authors please comment on whether there may have been a change in diagnosis process and/or accuracy over time that could be impacting on the study inferences?

Similarly, whether changes in individuals' capacity to treat infections over time might have an affect on the study findings?

"The analysis included 291,941 AD cases (median age at diagnosis: 76.2 years; male: 36 46.6%), 103,919 PD cases (74.3; 55.1%), and 10,161 ALS cases (69.3; 56.8%)."

Can the authors please comment on whether the analysed samples can be considered to be representative of the wider populations of interest?

"Individuals with multiple neurodegenerative diseases contributed to the analyses of different diseases."

Did the authors consider conducting a sensitivity analysis excluding individuals with multiple neurodegenerative diseases within each analysis?

"Five controls per case, individually matched by sex and year of birth, were randomly selected from the study base using incidence density sampling method."

A rigorous matching methodology has been applied by the authors.

"We then applied conditional logistic regression to estimate odds ratio (OR) and 95% confidence 135 interval (CI), as an estimate of the association."

The authors have applied technically appropriate modelling methods within the context of this research.

"Due to the concern about diagnostic delay, we used a lag time of 5 years, namely infections experienced during 5 years before the index date were excluded, in the main analysis. "

Can the authors please clarify how individuals with multiple cases of infection (< and > 5 years before diagnosis) were accounted for in the study? i.e. was the infection excluded if within five years of diagnosis, or the individual?

The authors have undertaken a thorough and extensive array of secondary and sensitivity analyses, which help to demonstrate the robustness of the study findings.

The Results are presented accurately and the main study limitations are transparently discussed.

*** Reviewer #2:

This is a case-control study nested within national Swedish registers examining the associations of hospital-treated infections with Alzheimer's disease, Parkinson's disease and amyotrophic lateral sclerosis. The study is interesting and timely and benefits from a very large dataset. I also commend the authors for 5- and 10-year lag analyses to reduce the risk of reverse causation and ascertainment bias. My suggestions to improve the study are as follows:

Major points:

Because this is an observational study, I would suggest using more cautious language in the title. For example, "associated with an increased risk" would be more appropriate than "increase the risk", which implies causality. The title would perhaps also be more informative if it mentioned ALS as well.

The methods mention the specificity of the register-based outcome event definitions. It would be fair to report their sensitivity as well.

Regarding Table 1: I would assume the characteristics were recorded at index date rather than baseline. Please clarify.

Please clarify when were each of the covariates recorded. I would say that appropriate times for recording them would be on index date or earlier but not later (to prevent the covariates from being affected by the outcome).

Missing information in education was included as a separate category for adjustments. This approach would be fine if missing data were rare, but in this case up to half of the participants had missing information on education, which may lead to confounded estimates.

Please clarify, how was the index date defined in the analysis of at least two outcome events. Was it that of the first or second event? Where are these findings reported?

In the Discussion, the authors write that assessing only hospital-treated infections underestimates true associations. I think overestimation is also possible as hospital-treated infections are obviously more severe than those not requiring hospital treatment. This thought is also supported by a previous study, in which the association of primary care-recorded infections with dementia was very weak (Muzambi et al. Lancet Healthy Longev 2021;2(7):e426-35).

The first sentence of the Discussion states that individuals with neurodegenerative diseases were more likely to experience hospital-treated infections. This is confusing, because only 2 of the 3 assessed neurodegenerative conditions were associated with infections.

Vascular, traumatic and alcohol-related brain damage are common aetiologies of early onset dementia (e.g., McMurtray et al. Dement Geriatr Cogn Disord. 2006;21(2):59-64). Thus, in the interpretation of the findings, it should be taken into account that the observed associations of early and middle age infections with dementia could at least in part be explained by confounding due to vascular risk factors, head traumas or alcohol or other substance problems. Could these factors also be taken into account in the analysis of the data?

Minor points:

Please clarify why were only those whose parents were born in Sweden included in the study.

The statistical analysis section states that ORs from nested case-control studies mimic relative risk estimates of the underlying cohort study. To my understanding, this depends on the sampling method used unless the outcome is rare.

Although most previous studies focused on specific infections, this is not true of all the previous studies cited in the manuscript, although the first sentence of the strengths and limitations section of the Discussion (p. 16) seems to imply so.

Discussion, p. 17, lines 317-319: I agree that the register-based approach does not capture all patients with neurodegenerative diseases. However, contrary to what the text implies, this cannot necessarily be deduced from the low positive predictive values, but is rather dependent on the sensitivity of the outcome ascertainment procedure.

The manuscript would benefit from language checking.

*** Reviewer #3:

Summary

This was a clearly-written paper that used a nested matched case control design in Swedish National Patient Register data to compare the likelihood of hospitalised infections occurring more than 5 years prior to diagnosis between 291,941 cases of AD, 103,919 of PD or 10,161 of ALS and up to 5 controls per case matched on age and sex. Prior hospital-treated infections were more common among AD cases (1.16, 95% C.I. 1.15-1.18) and PD cases (1.04, 95% C.I. 1.02-1.06) but not ALS cases (OR 0.97, 95% C.I. 0.92-1.03). Associations were stronger for younger patients, and those with multiple infection episodes.

Originality/ importance

Whether infections are a causal risk factor for neurodegenerative diseases is an important question with scope for major health impact. Existing longitudinal studies of clinically symptomatic infections and dementia have tended to focus on limited infection types e.g. sepsis or pneumonia1,2. There have been two recent large cohort studies from the UK and Finland, which showed positive associations between broad groups of medically-diagnosed infections and incident dementia risk3,4. This study extends findings to a different setting, with longer follow up and investigates two other neurodegenerative conditions, ALS and PD, as well as Alzheimer's disease.

Specific comments:

1. In the methods section on p6, high figures are given for the specificity of register-based definitions of the three outcomes. It would be helpful to include figures for other aspects of validity such as sensitivity and positive predictive value, especially as the potential for outcome misclassification is later suggested as a possible limitation.

2. The study time period begins in 1970. While this is a strength in terms of length of follow up, there have presumably been major changes in diagnostic and recording practices for exposures, outcomes and covariates over that time period. Creating time strata of 1970-1986 and 1987-2016 may not be sufficient to highlight differences in results in different time periods. Did you consider stratifying further? Although there was some discussion of data completeness over time, it would be useful to consider this issue further.

3. Exposure was hospitalised infections occurring at least five years prior to the index date. It was unclear how infections occurring between 5 and 0 years before index were treated, and whether this might lead to exposure misclassification

4. For the AD analysis, over half of participants had missing data on education. How was this dealt with in the analysis?

5. Models were adjusted for a relatively limited number of confounding factors, which might have led to residual confounding. While a lack of data on factors such as BMI, smoking, alcohol was mentioned as a limitation, other potential confounding factors such as ethnicity and socioeconomic status were also not included (the latter would be incompletely captured by education or area of residence). In addition, the rationale for the approach to clinical comorbidities was unclear: why create a Charlson comorbidity index rather than describe and adjust for individual comorbidities separately?

6. Health-seeking behaviour is likely to affect recording of both exposure and outcome. Was it possible to describe, or adjust for, frequency of consultations or healthcare contacts in the different exposure groups?

Minor comments:

1. The title should fully reflect the study - currently ALS is not mentioned.

2. Table 1 gave the proportion of the study population with any comorbidities, but it would be better to present Charlson comorbidity index scores (if this was the analytical approach taken)

3. The English language should be carefully checked throughout.

References

1. Muzambi R et al. Common Bacterial Infections and Risk of Dementia or Cognitive Decline: A Systematic Review. J Alzheimers Dis. 2020;76(4):1609-1626

2. Chu CS et al. Bacterial pneumonia and subsequent dementia risk: A nationwide cohort study. Brain Behav Immun. 2022 Apr 4;103:12-18

3. Muzambi R et al. Assessment of common infections and incident dementia using UK primary and secondary care data: a historical cohort study. Lancet Healthy Longev. 2021 Jul;2(7):e426-e435

4. Sipilä PN et al. Hospital-treated infectious diseases and the risk of dementia: a large, multicohort, observational study with a replication cohort. Lancet Infect Dis. 2021 Nov;21(11):1557-1567

*** Reviewer #4:

"Hospital-treated infections in early and mid-life increase the risk of Alzheimer's disease and Parkinson's disease" (PMEDICINE-D-22-01054R1)

The objective of this paper was to examine "the risk of three most common neurodegenerative diseases in relation to previous inpatient or outpatient episodes of hospital-treated infections." Swedish national data were used, with a follow-up from 1970 until 2016. The results show hospital-treated infections to be associated with a higher risk of Alzheimer's disease (AD) and Parkinson's Disease but not amyotrophic lateral sclerosis.

I have a number of concerns, listed below.

1. The statement "Infection is one of the most discussed potential non-genetic risk factors for neurodegenerative disease" (line 69-70) needs further elaboration as infection does not figure on most compilations of risk factors for neurodegenerative diseases, AD in particular.

2. In their review of the current status of knowledge the authors state that the role of "diagnostic delay" (line 78) has not been addressed in previous research. It isn't clear what this means, do they mean delay in diagnosis of neurodegenerative diseases? If this is the case, I did not find any analyses on diagnostic delay.

3. A major concern with this paper is what it adds to the recent paper by Sipila et al. Lancet Infect Dis 2021 (reference 17 in the manuscript). The measures of exposure and outcome, and conclusion are much the same.

4. The use of a lifecourse approach is promoted in the introduction but the authors chose to use a case-control design, this is a pity. Time to event analyses would have suited the data better; furthermore, the choice of controls retrospectively is never straightforward. It would be certainly better to retain everyone in the analyses so that people who die over the follow-up can be censored so that mortality does not bias findings.

5. There is no mention of period or cohort effects in the results; it is possible that these play a role, particularly for the exposure of interest.

6. In the primary analysis (lines 186-189) it is not clear if the results for men and women were different as statistical tests for difference are not provided.

7. Figure 3 shows that the association of infection and AD was confined to young-onset AD (under 60 years). These results are intriguing as the majority of AD cases in the population are late onset AD. Early onset AD is primarily driven by genetic factors and it is surprising to have 26% of cases in this study to be early onset AD. Please clarify the age at AD onset in this group; the median age at onset of AD in the total study population was 76.2 years (page 163).

8. It would be better to stratify results reported in Figure 4 as a function of early- and late-onset AD (after 65 years).

9. The first sentence of the discussion (line 230-231) would be better supported using time to event analyses and a wash-out period of 20 years.

***

Any attachments provided with reviews can be seen via the following link:

[LINK]

30 May 2022

Submitted filename: letter to editor and reviewer.docx

Click here for additional data file.

26 Jun 2022

Dear Dr. Sun,

Thank you very much for submitting your revised manuscript "Hospital-treated infections in early- and mid-life and risk of Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis: a nationwide nested case-control study in Sweden" (PMEDICINE-D-22-01054R2) for consideration at PLOS Medicine.

Your paper was discussed with our academic editor, and re-sent to its four independent reviewers. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

In light of these reviews, we will again be unable to accept the manuscript for publication in the journal in its current form, but we would like to invite you to submit a further revised version that addresses the reviewers' and editors' comments fully. You will recognize that we cannot make a decision about publication until we have seen the revised manuscript and your responses, and we expect to seek re-review by one or more of the reviewers.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We hope to receive your revised manuscript by Jul 15 2022 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

Please let me know if you have any questions, and we look forward to receiving your revised manuscript.

Sincerely,

Richard Turner, PhD

Senior editor, PLOS Medicine

rturner@plos.org

-----------------------------------------------------------

Requests from the editors:

We suggest addressing referee 4's comments by reporting some additional analyses, perhaps in the supplementary information.

Please reformat the 'Author summary' so that the 3 subsections each consists of around 3 bulleted points, each comprising 1-2 short sentences.

At line 91 and any other instances, please use the general style "60 years".

"Methods" at line 147.

At the end of the main text, please remove the information on funding from the Acknowledgements (in the event of publication this will appear in the article metadata, via entries in the submission form, and it does not need to be repeated here).

Please make "p<0.001" the smallest p value quoted throughout, unless there are specific statistical reasons to do otherwise.

Throughout the text, please move reference call-outs prior to punctuation, e.g., "... genetic causes [6,7]." at line 115.

Noting references 4 & 5, for example, please abbreviate journal names consistently.

Noting reference 44, please use the journal name abbreviation "PLoS ONE".

Comments from the reviewers:

*** Reviewer #1:

The authors have satisfactorily considered and responded to each comment in turn, adding to the analysis and amending the manuscript accordingly.

*** Reviewer #2:

The authors have done good job addressing my concerns. My few remaining concerns are as follows:

As a response to my question about the time of the recording of the covariates, the authors wrote that "[A]ll covariates were measured on the index date or earlier". I still struggle to understand what this means. Were there covariates for which data were not available up to the index date? If so, for what time period were the data available and used for defining each of these covariates? Please clarify in the manuscript the time periods used for the recording of the covariates.

In my initial review, I had concerns about missing data for education. I agree with the authors that conducting sensitivity analyses is the best way to address this issue. In the revised manuscript, they have conducted a sensitivity analysis in which they dropped the adjustment for education in the whole sample. This is not wrong, but does not take into account potential confounding from education. Therefore, I think the authors should conduct an additional sensitivity analysis among those with data available for education. Within this group, they should compare an analysis with adjustment for education to an analysis without adjustment for education. I think this would show whether adjustment for education would have any material effect on the results.

I also have a couple of additional points that do not affect the validity of the revised manuscript but would clarify reporting:

Please specify in the methods whether the chosen controls were free from all three assessed neurodegenerative diseases or just the single disease of concern in each analysis (or if some other definition was used).

p. 17, lines 306-309: The authors write that "Similar results were observed 2) when restricting the analyses to individuals without a family history of the disease, or 3) when defining neurodegenerative diseases through at least two hospital visits concerning the same disease." Probably I missed it, but I could not find these results. Please report them in the appendix. Please also specify whether you mean that the results were similar to those in the age-stratified analysis described on lines 303-306 on the same page or to those of the main analysis.

*** Reviewer #3:

Overall, I am satisfied that the authors have adequately addressed my comments. The new time-stratified analysis is helpful. I also agree with the approach to health seeking behaviour, which could indeed be acting a mediator in this context.

*** Reviewer #4:

[see attachment]

***

Any attachments provided with reviews can be seen via the following link:

[LINK]

Submitted filename: Plos Med June 2022.docx

Click here for additional data file.

2 Jul 2022

Submitted filename: letter to editor and reviewers_R2_final.docx

Click here for additional data file.

29 Jul 2022

Dear Dr. Sun,

Thank you very much for re-submitting your manuscript "Hospital-treated infections in early- and mid-life and risk of Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis: a nationwide nested case-control study in Sweden" (PMEDICINE-D-22-01054R3) for review by PLOS Medicine.

I have discussed the paper with my colleagues and the academic editor and it was also seen again by 3 reviewers.

Provided the remaining editorial and production issues are dealt with we are planning to accept the paper for publication in the journal.

The remaining issues that need to be addressed are listed at the end of this email. Any accompanying reviewer attachments can be seen via the link below. Please take these into account before resubmitting your manuscript:

[LINK]

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

In revising the manuscript for further consideration here, please ensure you address the specific points made by each reviewer and the editors. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments and the changes you have made in the manuscript. Please submit a clean version of the paper as the main article file. A version with changes marked must also be uploaded as a marked up manuscript file.

Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. If you haven't already, we ask that you provide a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract.

We expect to receive your revised manuscript within 1 week. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

We ask every co-author listed on the manuscript to fill in a contributing author statement. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript.

Please note, when your manuscript is accepted, an uncorrected proof of your manuscript will be published online ahead of the final version, unless you've already opted out via the online submission form. If, for any reason, you do not want an earlier version of your manuscript published online or are unsure if you have already indicated as such, please let the journal staff know immediately at plosmedicine@plos.org.

If you have any questions in the meantime, please contact me or the journal staff on plosmedicine@plos.org.

We look forward to receiving the revised manuscript by Aug 05 2022 11:59PM.

Sincerely,

Caitlin Moyer, Ph.D.

Associate Editor

PLOS Medicine

plosmedicine.org

------------------------------------------------------------

Requests from Editors:

From the academic editor:

1.The association is primarily being driven by the early-onset cases (<60 y). Once these are excluded, the OR is close to 1.0. Moreover, Fig 4 shows that 2 or more infections after 60 were in fact protective for AD ((OR 0.91 (0.89-0.93)), so to speak. It is true that the authors should look at the whole group first, but when they try to understand this deeper, the finding of early-onset stands out. This should be highlighted, and it is intriguing and counter-intuitive. The authors appear to de-emphasise this finding in favour of the whole group, but the message from the analysis appears to be that the finding is true only for early-onset cases.

Other editorial points:

2. Response to reviewers: Please address the remaining comments of Reviewer 4. Specifically, please discuss and present a figure/table of results broken down by age of diagnosis (i.e. older or younger than 60 years of age) for AD/PD/ ALS (sex, age at infection, type of infection, calendar period). Please also address the remaining issues raised by Reviewer 2.

3. Title: Please revise as follows and please capitalize the first word of the subtitle and please make this change in the manuscript submission system as well as the manuscript file. “Hospital-treated infections in early- and mid-life and risk of Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis: A nationwide nested case-control study in Sweden”

4. Competing interests: Please revise to “independent of the present study” in your statement.

5. Abstract: Line 33: Please mention the variables that are adjusted for in the analyses.

6. Abstract: Line 36-37: Please revise to: “A hospital-treated infection 5 or more years earlier was associated with…”

7. Abstract: Line 38-39: For the sake of comparison, please also present the results for those diagnosed at age 60 years or older. “The associations were primarily due to AD and PD diagnosed before 60 years”

8. Abstract: Line 44: Please delete “however” from the sentence.

9. Abstract: Line 51: We suggest “especially of AD” is not needed here as the association was apparent for both AD and PD.

10. Author summary: Line 89-90: We suggest revising to: “The underlying mechanisms for the link between infections and neurodegenerative disease may not be specific to certain pathogens or affected organs but possibly occur at the systemic level.”

11. Methods: Lines 164-168: Rather than referring to sensitivity/PPV as “varying” please comment on the relatively lower sensitivity/PPV in particular for AD and why this might be the case: “The register-based definitions of AD, PD, and ALS have been validated against gold-standard clinical workup, showing a high specificity but a varying sensitivity and positive predictive value for AD (99.7%, 32.5%, and 57%) [23], PD (>98%, 72%, and 71%) [24], and ALS (all >90%) [25]. Date of diagnosis was defined as the date of first hospital visit concerning the disease.

12. Methods: Lines 167-168: Please clarify if the date of diagnosis was the first date where AD/PD/ALS was recorded as the primary diagnosis in the register.

13. Methods: Lines 175-176: Thank you for including a copy of the pre-specified analysis plan as a supporting information file. In the analysis plan, you note that updates were made to the plan in response to peer review. Please make sure that any changes in the analysis-- including those made in response to peer review comments-- are identified as such in the Methods section of the paper, with rationale.

14. Results: Lines 275-277: Please reference the figure/table where these results are presented: “Positive associations for AD and PD were similarly observed for bacterial, viral, and other infections, as well as for CNS, gastrointestinal, and genitourinary infections.”

15. Results: Lines 290-291: “Stronger associations were also observed for AD ascertained before 65, compared with AD ascertained after 65” Please present the OR, 95% CI and p values in the text for this finding.

16. Results: Line 299-302: Please mention the table/figure where the analysis with number of infections is presented.

17. Results: Line 312-313: “Gastrointestinal infection was also associated with a higher risk of

313 AD (OR=1.35, 95%CI: 1.28-1.41, P < 0.001) and PD (OR=1.18, 95%CI: 1.06-1.32, P < 0.001).” Please clarify if this is among all individuals or those diagnosed before age 60 years.

18. Results: Lines 321-346: In the description of secondary analyses, we suggest removing the numbering from the paragraph (“1) A stronger association…2) when restricting the analyses…).

19. Results: Lines 325-339: Please include supporting information tables presenting the results of analyses restricted to those without a family history, neurodegenerative diseases defined by at least 2 hospital visits, from models not adjusted for education or restricted to those with complete education data, and when those with more than one neurodegenerative disease were excluded.

20. Results: As suggested by reviewer 4, please also present a table of results broken down by age of diagnosis (i.e. older or younger than 60 years of age) for AD/PD/ALS (sex, age at infection, type of infection, calendar period).

21. Discussion: Line 349-350, Please revise to: “In this nationwide study, we found that hospital-treated infections 5 years ago or earlier were associated with an increased risk of AD and PD, specifically among cases diagnosed before 60 years, but not ALS.”

22. Discussion: Line 355-357: Please revise to: “Though causality cannot be inferred from the study, the similar results for different infection types and sites might suggest that the potential underlying mechanisms of the observed associations are not specific to certain pathogens, and raise the possibility that systemic inflammation might play a role in brain health, an idea that is supported by evidence from a previous study reporting a relationship between hospital treated infection and vascular dementia and AD [17].” or similar.

23. Discussion: Line 385-387: Please revise this statement, as the paper cited actually reports finding no significant association between risk of PD more than 10 years after infection: “A recent study from Denmark reported an increased risk of PD more than 10 years after exposure to infections (OR=1.04, 95%CI: 0.98-1.10) [12].”

24. Discussion: Line 392-393: Please clarify if this refers to calendar year of diagnosis: “...the magnitude of the associations decreased over time.” Please comment on how the attenuation of the association in more recent calendar years might arise from better treatment for infection or improved diagnosis for AD/PD.

25. Discussion: Line 458-459: Please provide a reference: “The low sensitivity of AD diagnosis was likely attributable to misdiagnosis of AD as other dementias.”

26. Discussion: Line 470-473: Please comment on how the relatively young study population might impact the generalizability and interpretation of the findings, in light of the fact that associations were seen only among those diagnosed younger than 60 years of age.

27. Discussion: Line 478: Please use “substance misuse” here.

28. Discussion: Line 485-487: Please revise to: “In conclusion, our study suggested that individuals with hospital-treated infections, especially in those occurring in early- and mid-life, had an increased risk of developing AD and PD, attributable to cases diagnosed at young age.”

29. References: The journal abbreviation for reference 44 should be PLoS One.

30. STROBE checklist: For item 22, please refer to the location as “Funding” or similar.

31. Figure 1: Please define all abbreviations (AD/PD/ALS) in the legend.

32. Figure 2, Figure 3, Figure 4: Please define all abbreviations in the legend. Please add numbers of cases/controls and those with and without infection, as well as p values to the figure. Please note the factors adjusted for in the legends. Please also provide the results of unadjusted analyses (these can be added to the supporting information if preferred).

33. Table 2: Please add p values to the table. In the legend, please define all abbreviations used (AD/PD/CNS).

34. S3 -S10 Table: Please include p values. Please define AD, PD, ALS in the legends. Please also provide the unadjusted comparisons as well as the adjusted comparisons where applicable.

Comments from Reviewers:

Reviewer #1: The authors have satisfactorily considered and responded to each comment in turn, adding to the analysis and amending the manuscript accordingly.

Reviewer #2: The authors have responded to all my concerns, but their response needs to be clarified. When reporting the results of the sensitivity analyses suggested by me, they say that (p. 19, lines 333-336) "when restricting the analysis to individuals with complete data on education, consistent results were also noted between the models with or without adjustment for education (OR=1.10, 95%CI: 1.09- 1.12 for AD, P < 0.001; OR=1.03, 95%CI: 1.01-1.05 for PD, P < 0.001; and OR=0.98, 95%CI: 0.92-1.05 for ALS, P = 0.603)." Here, to support their text, they should show two ORs for each of the diseases (AD, PD, and ALS), one for the model adjusted for education and the other for the model not adjusted for education.

While reading the paper, I also noticed another point about clarifying the Discussion: On p. 21 (lines 372-376) the authors refer to an earlier study saying that "Another study reported a dose-response relationship between number of infectious diseases and dementia risk [17]. This study however did not explore the effect of age at infection and whether a dose-response relationship existed still after considering the influence of potential reverse causation due to the pre-clinical stage and diagnostic delay of neurodegenerative diseases." In a previous version of the manuscript they specified that the previous study did not consider the influence of potential reverse causation when looking at the dose-response relationship between dementia and AD, which was true. Now that the mention of AD has been omitted, the statement is not any more accurate as there was such an analysis on all-cause dementia in the previous study.

Reviewer #4: Hospital-treated infections in early- and mid-life and risk of Alzheimer's disease, Parkinson's disease, and amyotrophic lateral sclerosis: a nationwide nested case-control study in Sweden" (PMEDICINE-D-22-01054R3)

1. Apologies for repeating what I said earlier but the revised version of the manuscript re-mains problematic. The unstratified analyses make no sense because of the number of difference as a function of age at exposure, birth-year, age at onset of the neurodegenerative diseases, and sex. The results section of the abstract provides results that are incorrect, see below

"A hospital-treated infection 5 years ago or earlier was associated with an increased risk of AD (OR=1.16, 95%CI: 1.15-1.18, P < 0.001) and PD (OR=1.04, 95%CI: 1.02-1.06, P < 0.001)."

2. This is also the case in the following sections of the authors summary; these statements suggest that infection is a risk factor for AD. Early onset AD is different from AD and the distinc-tion is lost in these sentences.

What did the researchers find?

Individuals with repeated infections in early- and mid-life had the greatest risk increment of AD and PD.

What do these findings mean?

The underlying mechanisms for the link between infections and neurodegenerative disease

are likely not specific to certain pathogens or affected organs but occur at the systemic level.

Infectious events may be a trigger or amplifier of a pre-existing disease process, leading to

clinical onset of neurodegenerative disease at a relatively early age among individuals with

disease predisposition.

3. In the response to reviewers comments the authors state the following:

"We decided to report the unstratified analyses as the main results, because our primary hypothesis

was that there was an association between infections and risk of any AD/PD/ALS and these

analyses were a priori determined as the primary analyses. Although we did also hypothesize that

there might be variation in the associations by age at onset of neurodegenerative disease, etc., this

was a secondary hypothesis."

If this is the case then the authors ought to state clearly that their primary hypothesis was not verified and subsequent analyses showed an association for early onset AD/PD.

4. Following on from the previous comment, I don't quite understand the analyses reported in Figure 4. Surely, this ought to be stratified by age at onset of AD/PD.

Any attachments provided with reviews can be seen via the following link:

[LINK]

3 Aug 2022

Submitted filename: letter to editor and reviewers_R3.docx

Click here for additional data file.

15 Aug 2022

Dear Dr Sun,

On behalf of my colleagues and the Academic Editor, Perminder Singh Sachdev, I am pleased to inform you that we have agreed to publish your manuscript "Hospital-treated infections in early- and mid-life and risk of Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis: A nationwide nested case-control study in Sweden" (PMEDICINE-D-22-01054R4) in PLOS Medicine.

Before your manuscript can be formally accepted you will need to complete some formatting changes, which you will receive in a follow up email. Please be aware that it may take several days for you to receive this email; during this time no action is required by you. Once you have received these formatting requests, please note that your manuscript will not be scheduled for publication until you have made the required changes.

In the meantime, please log into Editorial Manager at http://www.editorialmanager.com/pmedicine/, click the "Update My Information" link at the top of the page, and update your user information to ensure an efficient production process.

Please also address the following editorial points:

-Results: Line 279-280: Please remove “CNS” from the list here, as no significant associations were found between CNS infection and PD, in contrast with AD, according to Table 2.

-Discussion: Line 339: Please revise to: “...hospital-treated infections 5 years or more prior to diagnosis were associated with increased risk…” or similar.

PRESS

We frequently collaborate with press offices. If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximise its impact. If the press office is planning to promote your findings, we would be grateful if they could coordinate with medicinepress@plos.org. If you have not yet opted out of the early version process, we ask that you notify us immediately of any press plans so that we may do so on your behalf.

We also ask that you take this opportunity to read our Embargo Policy regarding the discussion, promotion and media coverage of work that is yet to be published by PLOS. As your manuscript is not yet published, it is bound by the conditions of our Embargo Policy. Please be aware that this policy is in place both to ensure that any press coverage of your article is fully substantiated and to provide a direct link between such coverage and the published work. For full details of our Embargo Policy, please visit http://www.plos.org/about/media-inquiries/embargo-policy/.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Thank you again for submitting to PLOS Medicine. We look forward to publishing your paper.

Sincerely,

Caitlin Moyer, Ph.D.

Associate Editor

PLOS Medicine