Risk Factors for Atherothrombotic Disease and Lumbar Spinal Stenosis: Results from the PREFACE Study

Emilia Ruggiero¹, Simona Costanzo¹, Augusto Di Castelnuovo², Mariarosaria Persichillo¹, Simona Esposito¹, Sara Magnacca², Ettore Carpineta³, Chiara Cerletti¹, Marialaura Bonaccio¹, Maria Benedetta Donati¹, Sergio Paolini^3,4, Vincenzo Esposito^3,4, Giovanni de Gaetano¹, Gualtiero Innocenzi³ and Licia Lacoviello^1,5*
*Correspondence: Licia Lacoviello, Department of Epidemiology and Prevention, IRCCS NEUROMED, Pozzilli (IS), Italy, Email:

Keywords

Risk factors • Central obesity • Lumbar spinal stenosis • Case-control study • Socio-economic status

Introduction

Lumbar Spinal Stenosis (LSS) has become the most common indication for lumbar spine surgery in patients over 65 years [1], in part because of the increasing quality and availability of radiological imaging, in part for the higher need of mobility and flexibility in the aging population. Localized inflammation, triggered by nerve root compression [2] and at least two interdependent vascular mechanisms may contribute to the development of neurogenic claudication in LSS: reduced arterial blood flow resulting in ischemia and venous congestion with compression of the nerves and secondary perfusion deficiency [3]. Since ischemia may cause both pain and degeneration of the involved structures, atherothrombotic disease of the feeding arteries of the lumbar spine has attracted growing attention as one of the possible underlying mechanisms for LSS. Therefore, risk factors for atherothrombotic disease might represent an important target for primary and secondary prevention of the disease. Some studies have already observed an association between risk factors for cardiovascular disease and LSS [4]. Uesugi et al. reported a close association of diabetes and hypertension with LSS in 50 to 69-year-old Japanese patients [5]. Diabetes was also associated with symptomatic LSS in Japanese patients with moderate radiographic stenosis [6]. In a large cohort of Swedish construction workers, obese and overweight persons were at a higher risk of developing LSS [7]. A further analysis of the same cohort showed that tobacco smoking was, in a dose dependent manner, associated with an increased incidence of surgically treated LSS. However, all these studies considered one of few exposures at a time, without taking into consideration the potential presence of multiple risk factors or their interaction and/or possible confounding. The aim of our study was to evaluate the possible association of a large panel of risk factors for atherothrombotic disease, alone or in combination, with LSS in multivariable models.

Methods

Case patients

Between May 2016 and December 2018, 264 consecutive patients, aged ≥ 35 years, with symptomatic LSS, confirmed by imaging, eligible for surgery at the Neurosurgery Department of the IRCCS Neuromed, were enrolled in the PREFACE (PREdictive FACtors of risk and surgical outcomEs in lumbar spinal stenosis) study. Subjects with spine instability were excluded. Anthropometric measurements and administration of questionnaires were completed before surgery. The final sample consisted of 213 cases. The study was approved by the Ethical Committee of the IRCCS NEUROMED.

Control subjects

Controls were 426 subjects, without any reported spine degenerative disease or clinical sign of back disease, selected among participants of the Moli-sani Study, a large population-based cohort, of men and women (aged ≥ 35 years) randomly enrolled from the general population of the Molise region, where the IRCCS Neuromed is located [8]. For each case, we selected two controls matched by sex, age (± 6 months) and physical activity (low, moderate, intense).

Assessment of covariates

Physical activity level was assessed through the question “How was your leisure-time physical activity in the last six months?” with three possible answers (sedentary lifestyle, low active lifestyle or physically active lifestyle). Evaluation of covariates was conducted with identical procedures both in cases and in controls. Education was based on the highest qualification attained and was categorized as up to secondary (≤ 8), upper secondary (>8 ≤ 13), and post-secondary (>13). Present occupation was described through 5 groups: non-manual or manual occupation, retired, housewife and unemployed/unclassified. Former employment considered the previous occupation for retired; as a result subjects were classified as non-manual, manual, housewife and unclassified. Subjects were classified as never-smokers, current smokers or ex-smokers (having quitted for at least 1 year). Height and weight were measured, and Body Mass Index (BMI) was calculated as kg/m² and then grouped into 3 categories as normal (≤ 25), overweight (>25 to <30), or obese (≥ 30).

Moreover, waist circumference, in cm, was measured in the middle between the twelfth rib and the iliac crest and hip circumference, in cm, was measured around the buttocks. The waist-to-hip ratio was then calculated. Abdominal obesity was defined as waist-to-hip ratio ≥ 0.90 for men and ≥ 0.85 for women [9]. Hypertension, hypercholesterolemia and diabetes were defined by current pharmacological treatment. History of cardiovascular (angina, stroke and myocardial infarction) and peripheral artery disease included self-reported diagnosis. The study sample was also stratified as living in an urban or rural environment on the basis of the urbanization level as defined by the European Institute of Statistics (EUROSTAT definition) and obtained by the tool “Atlante Statistico dei Comuni” provided by the Italian National Institute of Statistics (www.istat.it) [10].

Statistical analysis

The main characteristics of cases and controls were reported as frequency and percentages or mean values and Standard Deviation (SD). Differences between cases and controls for categorical or continuous variables were evaluated by general linear models (PROC GENMOD and PROC GLM in SAS for categorical and continuous variables, respectively), taking into account the match for age, sex and physical activity. Multivariable conditional logistic regression analysis was used to quantify the association of potential risk factors or their combination with LSS status. Odds Ratios (ORs) and their 95% Confidence Intervals (CIs) were calculated using conditional-to-match (for age, sex and physical activity) logistic regression. Multivariable model included occupation, income, urban or rural residence, smoking habits, waist-to-hip ratio, physical activity, history of hypertension, diabetes or dyslipidaemia and peripheral artery disease as potential risk factors (Table 1). The multivariable model incorporated potential risk factors that had resulted associated with LSS status with p-value <0.20 in matched analysis.

An Atherothrombotic (AT) risk score was calculated, allocating 1 point for each risk factor: current smoking, abdominal obesity, hypertension, diabetes, dyslipidaemia. All the five risk factors contributed equally to the AT risk score which potentially ranged from 0 to 5. Three multivariable models were fitted. Model 1 included age, sex and physical activity; model 2 was additionally adjusted for occupation; model 3: as model 2 plus all the risk factors included into the score. Appropriate multiplicative terms for testing interaction between pair of risk factors were included in the model 2. The data analysis was generated using SAS/STAT software, Version 9.4 of the SAS System for Windows©2009. SAS Institute Inc. and SAS are registered trademarks of SAS Institute Inc., Cary, NC, USA.

Results

The baseline characteristics of cases and controls are shown in Table 1. In both cases and controls the mean age was 66.5 years (SD=9.3), 34.3% were women and 69.1% were subjects with a sedentary lifestyle. There was no difference between cases and controls regarding educational level, BMI distribution, or history of cardiovascular disease (Table 1). In univariable analysis, patients with LSS, matched for age, sex and physical activity with controls, were more likely to have manual occupation, lower income, to live in an urban area and to be smokers, along with a higher prevalence of hypertension, diabetes, dyslipidaemia and history of peripheral artery disease (Table 1). When all these conditions were included in a multivariable analysis, present manual occupation, current smoking, high waist-to-hip ratio and history of hypertension, diabetes and dyslipidaemia remained independently associated with higher odds of suffering of LSS (Model 2, Table 1). Table 2 reports AT risk score, its distribution in cases and controls and odds ratio for LSS.

AT risk score was higher in cases than in controls and was associated with an increased odd of LSS (OR: 2.22; CI 95%: 1.72- 2.86, Table 2, Model 2). Further adjustment of BMI did not modify the association between AT risk score and LSS (OR: 2.32; CI 95%: 1.78-3.02). Absence or one metabolic risk factors was observed in 38.7% of cases and 17.4% of controls, while 41.1% of cases showed three or more risk factor as compared to 22.7% of controls (Table 2). Risk of LSS increased progressively with increasing number of risk factors, with the highest category (> 3 components) associated with the greater LSS risk in comparison with individuals with zero components (OR: 13.04; 95% CI: 2.87-59.27; Table 2, Model 2 and Figure 1). Multiplicative terms of interaction included in Model 2 were all largely not statistically significant. In Model 3, we further adjusted for each component included into the score, and as a consequence the AT risk score completely abolished its association with LSS (Table 2, Model 3).

Discussion

A number of risk factors for LSS were evaluated in the context of a case-control design including 213 cases and 426 controls. Patients with LSS were more likely to report manual occupations and lower income, to live in an urban area, and to be smokers. This is in line with previous evidence indicating that severe LSS, assessed by MRI scanning, was associated with heavy manual work particularly in the factory/ construction industries [11] and tobacco smoking was associated with increased incidence of surgically treated LSS [7]. Among metabolic risk factors, diabetes was a major risk factor for LSS, as already seen by Asadian in Iranian subjects from a case-control study, where diabetes increased the risk of developing canal stenosis by 3.7 times [4]. BMI has been shown as an independent risk factor for symptomatic spinal stenosis in previous studies [12,13]. Our results found that central obesity rather than BMI was a powerful predictor of LSS. BMI has traditionally been used to measure obesity in epidemiological studies; however, waist-to-hip ratio more accurately describes the distribution of body fat and is more closely associated with morbidity and mortality [14]. Our results point out that visceral obesity, with its metabolic implications, rather than the simple increase in body weight, could play a role in the pathogenesis of LSS.

Taken altogether, the majority of previous studies took only into consideration one or few factors at a time, with limited power to control for possible confounding. Indeed, risk factors for atherosclerosis are often simultaneously present and multivariable models are necessary to dissect their independent effect on the disease. When a multivariable model was performed, including all the risk factors associated with LSS, manual occupation, waist-to-hip ratio, current smoking, and history of hypertension, diabetes and dyslipidaemia remained all associated with LSS, suggesting that they all contributed independently to the risk of LSS. This implies that the simultaneous presence of different risk factors can additively contribute to the risk or multiply their effect. To better explore this hypothesis, we elaborated a score of risk factors assessing the association of their combination on the risk of LSS in our study population. To date, very limited evidence was available on the association between combined factors and risk of LSS, with the exception of one study by Memtsoudis et al. indicating that patients with metabolic syndrome (actually a combination of risk factors) were more likely to undergo surgery for spinal stenosis [15].

Our results show that the AT risk score was higher in LSS patients than in the control group and was associated with LSS risk after adjustment for social status confounders. Each additional metabolic component doubled the odds of LSS risk in comparison with individuals with zero components. We also tested the possibility of interaction between risk factors by adjusting the multivariable model for the individual components of the AT risk score and by formally testing for interaction pairs of risk factors. In both cases, the hypothesis of additively was not rejected. To the best of our knowledge, this is the first study that has associated combination of AT risk factors to LSS. Although individual metabolic risk factors, such as abdominal obesity, hypertension or diabetes have been previously associated with LSS, very seldom the global effect of such disorders was distinguished from that of the increased body weight. Indeed, the effect of metabolic components could be merely attributed to an abnormal and altered load on the spine by BMI or body weight and to a relative reduction in muscle mass, which further increases strain on the lumbar spine [16-18].

Our study points out that besides direct biomechanical effect on cartilage and skeleton, central obesity acts through altered metabolism that induces changes in insulin sensitivity, levels of fatty acids, cytokines and growth factors. All these factors have the potential to alter the properties of bone matrix, ligament, synovia, and cartilage, and promote the development of spine and disc degeneration [19-26]. Furthermore, ischemic change of neural structures by these risk factors might contribute to the development of the cauda equina syndrome, a hypothesis supported by previous evidence [27]. Decreased muscle mass is indeed associated with insulin resistance, which further impairs the skeletal muscles and promotes systemic inflammation [28]. Adiponectin and leptin, hormones secreted by adipocytes function, determine low-grade inflammation that has been related to the progression of spondylosis [29,30]. In addition, a high serum concentration of free fatty acids increases systemic inflammation and induces development of osteoarthritis [31-33]. Finally, hyperlipidemiainduced atherosclerosis is proposed as a cause of disc degeneration and ischemic pain [34,35]. We tried to dissociate the effect of obesityrelated load from that of obesity associated-metabolic disorder by adjusting for BMI the association between AT risk score and LSS, but no change was observed. Additionally, in our sample, BMI was not associated with LSS in multivariable analysis taking into account metabolic features.

Conclusion

Risk factors for atherothrombosis are independently and additively associated with LSS. Central obesity reduction, control of metabolic factors and smoking cessation should be prominent measures in prevention and management strategies of LSS.

Acknowledgement

PREFACE Study Investigators

Principal Investigators: Licia Iacoviello, MD, PhD, (IRCCS Neuromed, Pozzilli and Università dell’Insubria, Varese, Italy) and Gualtiero Innocenzi, MD (IRCCS Neuromed, Pozzilli. Italy).

Department of Epidemiology and Prevention, IRCCS Neuromed, Pozzilli, Italy: Marialaura Bonaccio, Chiara Cerletti, Simona Costanzo, Amalia De Curtis, Giovanni de Gaetano, Maria Benedetta Donati, Simona Esposito, Mariarosaria Persichillo.

Department of Neurosurgery, IRCCS Neuromed, Pozzilli, Italy: Vincenzo Esposito, Sergio Paolini, Giovanni Cardarelli, Ettore Carpineta, Grazia Centore, Mariacristina Cotugno, Marianna Damiano, Michelangelo De Angelis, Arianna de Gregorio, Alessandro D’Elia, Maria Antonietta Di Santo, Cristina Mancarella, Giovanna Santangelo, Piera Sciarra, Rocco Severino, Tommaso Vangelista.

Recruitment staff: Emilia Ruggiero (Coordinator), Mariacristina Cotugno, Simona Esposito and Giovanna Santangelo.

Data management: Simona Costanzo, Emilia Ruggiero and Marco Olivieri* (*Università del Molise, Campobasso, Italy).

Data analysis: Simona Costanzo, Emilia Ruggiero and Augusto Di Castelnuovo§ (§Mediterranea Cardiocentro, Napoli. Italy).

Biobank and biomedical analyses: Amalia De Curtis, Emilia Ruggiero, Simona Esposito, Marianna Storto^ (^Department of Analysis Lab Diagnostics, IRCCS Neuromed) and Sara Magnacca§ (§Mediterranea Cardiocentro, Napoli. Italy).

Informatics: Marco Olivieri* (*Università del Molise, Campobasso, Italy).

Strengths and Limitations

The novelty of our study is the simultaneous evaluation of several risk factors for LSS and the possibility to account for confounding and interactions. There are however some limitations to this study: it is possible that risk factors, occurring infrequently, may not have been identified. Moreover, the retrospective nature of a case-control study and the potential for recall bias might limit its clinical relevance.

References

1. Jeffrey N. Katz, Gerold Stucki, Stephen J. Lipson and Anne H. Fossel, et al. “Predictors of Surgical Outcome in Degenerative Lumbar Spinal Stenosis.” Spine 24 (1999):  2229-2233.
2. James D. Kang, Helga I. Georgescu, Lori McIntyre-Larkin and Maja Stefanovic-Racic, et al. “Herniated Lumbar Intervertebral Discs Spontaneously Produce Matrix Metalloproteinases, Nitric Oxide, Interleukin-6, and Prostaglandin E2.” Spine 21 (1996):  271-277.
3. Kauppila, Lui. “Atherosclerosis and Disc Degeneration/low-back pain:  A Systematic Review.” Eur J Vasc Endovasc Surg 37 (2009):  661-670.
4. Leila Asadian, Kaveh Haddadi, Mohsen Aarabi and Amirhossein Zare. “A New Risk Factor for Lumbar Spinal Stenosis:  A Case-Control Study.” Clin Med Insights Endocrinol Diabetes 9 (2016):  1-5.
5. Kazuhide Uesugi, Miho Sekiguchi, Shin-ichi Kikuchi and Shin-ichi Konno. “Relationship between Lumbar Spinal Stenosis and Lifestyle-related Disorders:  A Cross-sectional Multicenter Observational Study.” Spine 38 (2013):  540-545.
6. Takahiro Maeda, Hiroshi Hashizume, Noriko Yoshimura and Hiroyuki Oka, et al. “Factors Associated with Lumbar Spinal Stenosis in a Large-scale, Population-based Cohort:  The Wakayama Spine Study.” PLoS One 13 (2018):  208.
7. Björn Knutsson, Sebastian Mukka, Jens Wahlström and Bengt Järvholm, et al. “The Association between Tobacco Smoking and Surgical Intervention for Lumbar Spinal Stenosis:  Cohort Study of 331,941 Workers.” Spine J 18 (2018):  1313-1317.
8. Augusto Di Castelnuovo, Simona Costanzo, Mariarosaria Persichillo and Marco Olivieri, et al. “Distribution of Short and Lifetime Risks for Cardiovascular Disease in Italians.” Eur J Prev Cardiol 19 (2012):  723-730.
9. Cino, Nishida. “Body Fat Distribution and Non-communicable Diseases in Populations:  Overview of the 2008 WHO Expert Consultation on Waist Circumference and Waist-Hip Ratio.” Euro J Clin Nutrit 64 (2010):  2-5.
10. Luca, Salvati. “Proposal for a Statistical-Economic Assessment of Soil Degradation:  A National Scale Approach.” RIV 56 (2014):  312-343.
11. Yuyu Ishimoto, Cyrus Cooper, Georgia Ntani and Hiroshi Yamada, et al. “Factory and Construction Work is Associated with an Increased Risk of Severe Lumbar Spinal Stenosis on MRI:  A Case Control Analysis within the Wakayama Spine Study.” Am J Ind Med 62 (2019):  430-438.
12. Björn Knutsson, Bengt Sandén, Göran Sjödén and Bengt Järvholm, et al. “Body Mass Index and Risk for Clinical Lumbar Spinal Stenosis:  A Cohort Study.” Spine 40 (2015):  1451-1456.
13. Ryan Garcia, Patrick Messerschmitt, Christopher Furey and Henry Bohlman. “Weight Loss in Overweight and Obese Patients following Successful Lumbar Decompression.” J Bone Joint Surg Am 90 (2008):  742-747.
14. Huxley, Rule. “Body Mass Index, Waist Circumference and Waist:  Hip Ratio as Predictors of Cardiovascular Risk:  A Review of the Literature.” Eur J Clin Nutr 64 (2010):  16-22.
15. Stavros G. Memtsoudis, Meghan Kirksey, Yan Ma and Ya Lin Chiu, et al. “Metabolic Syndrome and Lumbar Spine Fusion Surgery:  Epidemiology and Perioperative Outcomes.” Spine 37 (2012):  989-995.
16. Kap-Soo Han, Antonius Rohlmann, Thomas Zander and William R. Taylor. “Lumbar Spinal Loads Vary with Body Height and Weight.” Med Eng Phys 35 (2013):  969-977.
17. Heather K. Vincent, Kendrick Heywood, Jacob Connelly and Robert W. Hurley. “Obesity and Weight Loss in the Treatment and Prevention of Osteoarthritis.” PMR 4 (2012):  59-67.
18. Heather K. Vincent, Sara N. Raiser and Kevin R. Vincent. “The Aging Musculoskeletal System and Obesity-related Considerations with Exercise.” Age Res Rev 11 (2012):  361-373.
19. Francis Berenbaum and Jérémie Sellam. “Obesity and Osteoarthritis:  What are the Links?” Joint Bone Spine 75 (2008):  667-668.
20. Farshid, Guilak. “Biomechanical Factors in Osteoarthritis.” Best Pract Res Clin Rheumatol 25 (2011):  815-823.
21. Leonid Kalichman, Ali Guermazi, Ling Li and David J. Hunter. “Association between Age, Sex, BMI and CT-evaluated Spinal Degeneration Features.” J Back Musculoskelet Rehabil 22 (2009):  189-195.
22. Hirofumi Kosaka, Koichi Sairyo, Ashok Biyani and Douglas Leaman, et al. “Pathomechanism of Loss of Elasticity and Hypertrophy of Lumbar Ligamentum Flavum in Elderly Patients with Lumbar Spinal Canal Stenosis.” Spine 32 (2007):  2805-2811.
23. Like, Manto. “Disc Degeneration of the Llumbar Spine in Relation to Overweight.” Int J Obes 29 (2005):  903-908.
24. Piscoya, James. “The Influence of Mechanical Compression on the Induction of Osteoarthritis-related Biomarkers in Articular Cartilage Explants.” Osteoarthrit Cart 13 (2005):  1092-1099.
25. Koichi Sairyo, Ashok Biyani, Vijay Goel and Douglas Leaman, et al. “Pathomechanism of Ligamentum Flavum Hypertrophy:  A Multidisciplinary Investigation based on Clinical, Biomechanical, Histologic, and Biologic Assessments.” Spine 30 (2005):  2649-2656.
26. Koichi Sairyo, Ashok Biyani, Vijay K. Goel and Douglas W. Leaman, et al. “Lumbar Ligamentum Flavum Hypertrophy is due to Accumulation of Inflammation-related Scar Tissue.” Spine 32 (2007):  340-347.
27. Judita Orendáčová, Dáša Čı́žková, Jozef Kafka and Nadežda Lukáčová. “Cauda Equina Syndrome.” Prog Neurobiol 64 (2001):  613-637.
28. Paresh Dandona, Ahmad Aljada, Ajay Chaudhuri and Priya Mohanty, et al. “Metabolic Syndrome:  A Comprehensive Perspective based on Interactions between Obesity, Diabetes, and Inflammation.” Circulat 111 (2005):  1448-1454.
29. Dumond Hanne, Nathalie Presle, Bernard Terlain and Didier Mainard, et al. “Evidence for a Key Role of Leptin in Osteoarthritis.” Arthritis Rheum 48 (2003):  3118-3129.
30. Timothy M. Griffin, Janet L. Huebner, Virginia B. Kraus and Farshid Guilak. “Extreme Obesity due to Impaired Leptin Signaling in Mice does not Cause Knee Osteoarthritis.” Arthritis Rheum 60 (2009):  2935-2944.
31. Brunner, Aama. “High Dietary Fat and the Development of Osteoarthritis in a Rabbit Model.” Osteoarthr Cartil 20 (2012):  584-592.
32. Paresh Dandona, Ahmad Aljada, Ajay Chaudhuri and Priya Mohanty, et al. “Metabolic Syndrome:  A Comprehensive Perspective based on Interactions between Obesity, Diabetes, and Inflammation.” Circulat 111 (2005):  1448-1454.
33. Timothy M. Griffin, Janet L. Huebner, Virginia B. Kraus and Zhen Yan, et al. “Induction of Osteoarthritis and Metabolic Inflammation by a Very High-fat Diet in Mice:  Effects of Short-term Exercise.” Arthritis Rheum 64 (2012):  443-453.
34. Päivi Leino-Arjas, Leena Kaila-Kangas, Svetlana Solovieva and Hilkka Riihimäki, et al. “Serum Lipids and Low Back Pain:  An Association:  A Follow-up Study of a Working Population Sample.” Spine 31 (2006):  1032-1037.
35. Päivi Leino-Arjas, Leena Kauppila, Leena Kaila-Kangas and Rahman Shir, et al. “Serum Lipids in Relation to Sciatica among Finns.” Atheroscler 197 (2008):  43-49.