Associations Between Symptoms and Quality of Life in Lumbar Degenerative Disease and Cognitive Dysfunction in a Japanese Community: A Cross-sectional Study

Background: Lumbar degenerative disease (LDD) and dementia are increasing in super-aging societies and are both related to physical dysfunction and pain. However, the relationship between these diseases remains unclear. This cross-sectional study aimed to investigate the rate of comorbidity of lumbar spinal stenosis (LSS) and mild cognitive impairment (MCI) by age and sex and clarify the association between LSS presence, LDD symptoms, quality of life (QOL) related to low back pain, and cognitive impairment in a Japanese population. Methods: We enrolled 1097 participants (men 437; women 660; mean age 54.6 years) from a medical checkup program. LSS was diagnosed using a self-administered questionnaire, and LDD symptoms were evaluated using visual analog scale (VAS: low back pain, pain, numbness of the lower limb). QOL related to low back pain was evaluated using the Japanese Orthopedic Association Back-Pain Evaluation Questionnaire (JOABPEQ: pain, lumbar function, and gait function), and Kellgren-Lawrence grading was performed to evaluate lumbar degeneration using lateral radiographs of the lumbar spine. Cognitive function was measured using the Mini Mental State Examination (MMSE), and we dened MCI as a summary score of MMSE £27. Logistic and multiple liner regression analyses were performed to analyze the association between MCI, summary score of MMSE and LSS factors. Results: The comorbidity rate of MCI and LSS was 0.9% in men and 0.7% in women. It was 2.1% in those aged ³65 years in both sexes, and this rate increased with age. Lumbar function in JOABPEQ (OR 0.979, 95% CI 0.961 to 0.998) and lower limb pain in VAS (OR 1.020, 95% CI 1.002 to 1.039) were associated with MCI in men. The presence of LSS, lumbar function, and gait function in JOABPEQ were associated with MMSE in both sexes, while lower limb numbness and pain in VAS were associated with MMSE in men. Conclusion: The comorbidity rate as mean ± SD or as n (%). MCI, mild cognitive impairment; LSS, lumbar spinal canal stenosis; JOABPEQ, Japan Orthopedics Association Back Pain Evaluation Questionnaires; VAS, visual analog scale; KL, Kellgren-Lawrence grade; MMSE, Mini Mental State Examination; MCI, mild cognitive impairment; DM, diabetes mellitus; HT, hyper We showed the association between the factors of LDD and MMSE using multiple linear regression analysis with MMSE as the dependent variable. The presence of LSS (β=-0.214, 95% CI -3.018 to -1.344), lumbar function and gait function of JOABPEQ (lumbar function: β=0.161, 95% CI 0.008 to 0.028; gait function: β=0.106, 95% CI 0.002 to 0.022) and VAS of lower limb symptoms (pain of lower limb: β=-0.141, 95% CI -0.026 to -0.007; numbness of lower limb: β=-0.105, 95% CI -0.022 to -0.002) were related to MMSE signicantly in men. The presence of LSS (β=-0.093, 95% CI -1.211 to -0.168) and two domains of JOABPEQ (lumbar function: β=0.164, 95% CI 0.007 to 0.018; gait function: β=0.149, 95% CI 0.005 to 0.015) were associated with MMSE signicantly in women

dementia. Several reports have shown that cognitive impairment is related with pain and physical dysfunctions [7][8][9], which are typical symptoms of musculoskeletal disease. Cognitive function certainly decays with age, thereby increasing the prevalence of MCI in the future.
Previous studies investigating the association between cognitive function and musculoskeletal diseases reported that MCI was related with the occurrence of knee osteoarthritis [10] and that delayed union of osteoporotic vertebral fracture decreased cognitive function [11]. However, the association between LDD and cognitive function remains unclear, although both LDD and cognitive impairment are related to pain and physical dysfunction. Hence, clarifying the relationship between LDD and cognitive impairment may be important for improving ADL in elderly people. This study aimed to investigate the rate of comorbidity of MCI and LSS by age and sex, and to clarify the relationship between MCI/MMSE score and the risk factors of LDD-LSS presence, LDD symptom, QOL score depending on low back pain and radiographical lumbar degeneration in a community-dwelling Japanese population.

Participants
Our analysis was based on data collected from a medical checkup program in 2016. In brief, this program was initiated in 2005, and it was conducted over a 10-year period. In addition to orthopedic surgeons, gynecologists, urologists, endocrine physicians, cardiologist physicians, gastroenterologists, neurologists, and otolaryngologists were involved in this project. As one aspect of the multiple-focused check, we collected questionnaires and radiographic images related to musculoskeletal disorder.
For this cross-sectional study, 1149 participants were enrolled. Participants were asked to complete some self-administered questionnaires assessing their daily habits, medical histories, and LDD. They also underwent a cognitive screening test and a lateral lumbar radiograph in the neutral position. We excluded those who did not answer the questionnaire entirely, did not undergo cognitive screening test and radiographic examination, or had medical histories of lumbar spine surgeries. Of all participants, 52 were excluded (19 men, 33 women). Of those excluded, 36 (13 men, 23 women) did not answer questionnaires entirely, 13 (4 men, 9 women) did not undergo radiography, and 3 (2 men, 1 woman) had previous lumbar surgery. Finally, 1097 people (20-93 years old, 437 men, 660 women) were included in this study. The mean ages were 53.6 years old in men and 55.3 years old in women. All participants were provided a thorough explanation that the data collected would be analyzed, and gave written informed consent.

Diagnosis of LSS
LSS was detected using self-administered questionnaires consisting 15 questions associated with neurological disorders. We diagnosed LSS when the scores were 13 points or greater (sensitivity: 84%, speci city: 78%) [12].

Evaluation of cognitive function
Cognitive function was measured using the Mini Mental State Examination (MMSE) [13]. This is a 30item cognitive screening test that measures orientation, registration, short-term memory, attention and concentration, and language and construction capacity. The full score of MMSE is 30 points, with 0 being the worst cognitive function. We de ned the participant as having MCI when MMSE was 27 points or less, because this cut-off value had the greatest sensitivity (66.3%) and speci city (72.9%) of diagnosing MCI in a previous report [14]. We de ned the participants whose MMSE was score 27 points or less as MCI+ group, and the participants whose MMSE score was over 27 points as MCI-group.
Assessment of symptoms of LDD, and quality of life associated with low back pain The severities of the symptoms of LDD were evaluated with the visual analogue scale (VAS), for the most severe low back pain, lower limb pain and numbness during the past three months. The most severe score is 100 mm, and 0 mm means no symptoms.
Deterioration of quality of life (QOL) due to low back pain was evaluated using the Japanese Orthopedics Association Back Pain Evaluation Questionnaire (JOABPEQ) [15]. There have been several reports that JOABPEQ was useful for evaluating QOL in patients with lumbar disc herniation, LSS, and other lumbar diseases [16][17][18]. It consists of ve domains: pain, gait function, lumbar function, social life, and mentality, and each domain is scored out of 100 points. Three of the ve domains: pain, gait function and lumbar function were used for analysis in this study. Research assistants supported the participants who could not answer these questionnaires by themselves.

Measurement of lumbar degeneration on radiography
Lateral radiographs of the lumbar spine were taken, with the participants standing naturally, with their forearms crossed and hands on the chest. Radiographs were evaluated by a single orthopedic surgeon (KK) using Kellgren-Lawrence grading (KL) [19] in each intervertebral level (L1/2, L2/3, L3/4, L4/5, and L5/S1). To determine the severities of lumbar degeneration, the values of KL were summed from L1/2 to L5/S. According to this summed value, 0 corresponded to a normal lumbar spine and 20 expressed the most severe degenerative lumbar spine [20].

Medical histories and daily habits
All participants provided data related to their medical histories and daily habits which were previously reported as related to cognitive impairment [21][22][23][24][25]. Medical histories related to diabetes mellitus (DM), hypertension (HT), and depression were collected. We collected data pertaining to their duration of education (6 to 20 years), daily smoking habits (0, ex-smoker or never smoked; 1, current smoker), alcohol consumption (0, ex-drinker or never drank or social drinker; 1, habitual drinker) and exercise (0, no exercise habit; 1, having exercise habits over 2 times during one week).

Statistical analyses
Descriptive statistics were used to investigate the prevalence of MCI and LSS by age and sex. To compare characteristics between MCI+ and MCI-groups, we used the Mann-Whitney U test with the presence of MCI as the dependent variable and age, BMI, VAS, JOABPEQ, summed KL, and education periods as independent variables. To analyze the correlation between MCI and LDD parameters, logistic regression analysis was performed with MCI as the dependent variable and the presence of LSS, VAS, JOABPEQ and summed KL grades as the independent variables. For adjusting for age, duration of education, DM, HT, depression, smoking, alcohol consumption, and exercise, they were also included as independent variables. To analyze the correlation between MMSE and LDD parameters, multiple linear regression analysis was performed with MMSE as the dependent variable and the presence of LSS, VAS, JOABPEQ, and summed KL grades as the independent variables. All statistical tests were performed using SPSS ver. 22.0 (SPSS Inc., Chicago, IL, USA), and statistical signi cant was set at 0.05.

Prevalence of MCI and LSS by age and sex
In men, 48 individuals (11%) had MCI. The prevalence of MCI increased with age and was 29% (35/123) in those aged ³65 years. Eleven individuals (2.5%) were diagnosed with LSS and the prevalence of LSS also increased with age. In women, 45 individuals (6.8%) had MCI. The prevalence increased with age and was 18% (37/211) in those aged ³65 years. Twenty individuals (3.0%) were diagnosed with LSS. The comorbidity rate of MCI and LSS was 0.8% in all participants, 0.9% in men and 0.7% in women ( Table 1). The comorbidity rate was 2.1% (7/334) in those aged ³65 years, and increased with age. The rate of MCI was 29% (9/31) in participants with LSS in both sexes, 36% (4/11) in men, and 25% (5/20) in women. The rate of LSS was 9.8% (9/92) in participants with MCI in both sexes, 8.5% (4/47) in men, and 11% (5/45) in women.

Comparison of characteristics between MCI+ and MCIgroups
We investigated the differences of parameters between the group with MCI and that without MCI using the Mann-Whitney U test. Age and the prevalence of LSS and KL were signi cantly higher in the MCI+ group than in the MCI-group in both sexes. In men, lumbar function and gait function of JOABPEQ were lower in the MCI+ than in the MCI-group. In women, all three domains of JOABPEQ were lower in the MCI+ group than in the MCI-group, and pain and numbness of lower limb of VAS were higher in the MCI+ group than in the MCI-group (Table 2). The association between presence of LSS, symptoms of LDD, QOL related to LDD and MCI We showed the association between LDD factors (presence of LSS, symptoms related to LDD, and QOL score related to LDD) and MCI using logistic regression analysis with MCI as the dependent variable. In men, lumbar function of JOABPEQ (OR 0.979, 95% CI 0.961 to 0.998) and pain of lower limb of VAS (OR 1.020, 95% CI 1002 to 1.039) were signi cantly associated with MCI. In women, there was no signi cant relationship between MCI and any factors related to LDD (Table 3).

Summary of the current study
This population-based study was conducted to clarify the relationship between LDD and cognitive impairment. Initially, we found that the rate of comorbidity of MCI and LSS was 2.1% in those aged ³65 years. Second, we analyzed the relationship between LSS presence, LDD symptoms, deterioration of QOL due to low back pain, radiographical lumbar degenerations, and cognitive impairment. Lower limb pain and deterioration of QOL due to low back pain were signi cantly associated with MCI in men. LSS presence, and deterioration of QOL due to low back pain were signi cantly related to MMSE in both sexes, and pain and numbness of the lower limb were signi cantly associated with MMSE in men.

Prevalence of MCI and LSS, comorbidity of MCI and LSS
The prevalence of MCI and LSS in general populations have been reported previously. The prevalence of MCI was 3.0% to 42% [26,27], and had a strong association with aging [28]. Here, the prevalence of MCI was 8.5%, and had relation with aging. The range of prevalence of MCI was great in this and previous population studies, one reason being that the diagnostic tools used for MCI were different: such as Clinical Dementia Rating, MMSE, Montreal Cognitive Assessment, Psychogeriatric Assessment Scale, or Wechsler Memory Scale-Revised. We de ned MCI as a summary score of MMSE £ 27 in this study. MMSE is a useful screening tool for evaluating cognitive function. Thus, this study's results may be useful as base data for screening for cognitive impairment in a medical checkup program. Moreover, the ageranges of the subjects were different, and major cut-off age was 65 years old in previous studies. We collected data from the participants whose age-range was from 20 years old to 93 years old in this study.
Hence, the prevalence of MCI we showed would be reasonable compared with past reports.
The prevalence of LSS was 10.1% in men and 8.9% in women and was associated with age in the ROAD study, a nationwide study of the Japanese population [29]. In that study, LSS was diagnosed using lumbar MRI, medical histories, and physical test performed by an orthopedic surgeon, and based on North American Spine Society guideline. The current study showed that the prevalence of LSS was 2.5% in men and 3.0% in women and increased with age. The prevalence of LSS was lower in our study than that in the ROAD study because the diagnostic tool used for LSS was different, wherein we used a selfadministered diagnostic support tool for LSS. However, the diagnostic support tool we used had good sensitivity (84%) and speci city (78%) for diagnosing LSS [12]. Thus, LSS prevalence in the current study was meaningful as a base data for screening. LSS prevalence increases with age according to past and current studies [29].
To the best of our knowledge, there has been no report on the comorbidity of MCI and LSS to date. The comorbidity rates of MCI and LSS were 0.8% and 2.1%, respectively in those aged ³65 years in our study.
The rate of MCI was 29% in participants with LSS, and the rate of LSS was 9.8% in participants with MCI in both sexes. The prevalence of MCI and LSS has been proven to increase with age, and the comorbidity rate of MCI and LSS also increased with age in this study. We showed a high rate of comorbidity of MCI and LSS, and one third of the people with LSS had MCI in a Japanese dwelling population. This study was held as part of a medical checkup program, and almost all subjects would be interested in health. Thus, the comorbidity rate of MCI and LSS may increase if people who are not interested in health and who cannot go outside because of decaying mobility capability were enrolled as subjects in a similar study. The percentage of elderly persons in Japan will highly increase in the future, suggesting that the comorbidity rate of MCI and LSS will correspondingly further increase. Hence, when we examine and treat people with LSS, we should always consider the cognitive impairment as a potential comorbidity.

The relationship between LDD and cognitive function
Several reports were showed the links between musculoskeletal disease and cognitive function previously. Baseline MMSE summary score and the prevalence of MCI were signi cantly associated with the incidence of knee osteoarthritis [10]. Delayed union of vertebral fracture decreased MMSE summary score in a longitudinal study [11]. However, to the best of our knowledge, there has been no report on LDD and cognitive impairment. This study is the rst report showing that the presence of LSS, symptoms of LDD, and deterioration of QOL due to low back pain are associated with MCI and MMSE summary score.
Here, pain and numbness of lower limbs was associated with MCI and MMSE in men. The relationship between pain and cognitive impairment were also reported previously. Pain was a serious symptom secondary to neuropsychiatric symptoms in patients with dementia [30] and approximately 50% of people with dementia experienced pain regularly [31]. Chronic and neuropathic pain are particularly associated with cognitive impairment [7,8]. A past report showed that severe pain led to cognitive impairment in a longitudinal study [3]. We showed that neuropathic pain and numbness due to LDD were associated with cognitive dysfunction. A few reports showed the difference of pain threshold between men and women. Calvo-Perxas [32] showed gender differences in depression and pain, and suggested that mild to severe pain showed an association with depression in men, although only severe pain was associated with depression in women. They suggested that the frequency of pain, rather than severity of pain, was related with the deterioration of the central nervous system in men. In our report, the symptoms of the lower limbs might be associated with cognitive function in only men because of gender differences in the central sensitization of pain.
Several reports have showed the relations between physical dysfunction and cognitive impairment. Gait speed, standing balance, stand-up time and leg strength declined more in cognitive impairment subjects compared with healthy subjects [1,33,34], and these physical functions decayed prior to cognitive decline [9]. In the current study, deteriorations of QOL, which were related with lumbar function and gait function, were associated with the prevalence of MCI in men and decline of MMSE scores in both sexes.
Physical dysfunction may cause cognitive dysfunction through deterioration of QOL in people with LDD. The effect of intervention of pain and physical dysfunction on cognitive function is controversial. We should do an interventional study on whether LDD treatment also improves cognitive function.

Limitations
There were several limitations to our study. First, we used only the MMSE for evaluating cognitive function. Although MMSE is a standard measure of cognitive function, MMSE could evaluate only global cognitive function. We should study the relationships between LDD and speci c cognitive functions such as memory and language, among others. Second, although more than 1000 participants were included in this study, the study population may not be representative of the general population because participants were recruited from only one area of Japan. Further, we could not clarify the causal relationship between LDD and cognitive impairment because this was a cross-sectional study. Hence, a longitudinal study investigation, the causal relationship between LDD and cognitive function is necessary in the future.
Nevertheless, this is the rst study to have investigated the relationship between LDD and cognitive impairment in a Japanese population to the best of our knowledge, and our results provide valuable information to improve QOL of elderly people in a super-aging society. This health promotion project started from 2005 and provides continuous data, enabling longitudinal study to clarify further relationships between LDD and cognitive impairment.

Conclusions
Our results indicated that the comorbidity rate of MCI and LSS was 0.8% in all participants, and 2.1% in those aged ³65 years, and this comorbidity rate increased along with age. Deterioration of QOL related to lumbar function and pain of lower limb were associated with MCI in men. Deterioration of QOL related to lumbar and gait function, and LSS presence were associated with summary score of MMSE in both sexes, while pain and numbness of lower limb were associated with summary score of MMSE in men. Treatment of LDD may be useful in bringing improvement of cognitive function.