Risk of Pneumonia After Vertebral Compression Fracture in Women with Low Bone Density a Population-Based Study

EPIDEMIOLOGY

Risk of Pneumonia After Vertebral Compression Fracture in Women With Low Bone Density A Population-Based Study

Byunghyun Kim, MD, Joonghee Kim, MD, You Hwan Jo, MD, Si-Hyuck Kang, MD,y Yeon Joo Lee, MD,z Jae Hyuk Lee, MD, Je Eun Hwang, MD, Min Ji Park, MD, and Sihyung Lee, MD

increased risk of pneumonia (adjusted hazard ratio, 2.13; 95% Study Design. Retrospective large population-based study. confidence interval 1.47–3.08). Objective. The aim of this study is to elucidate the association Conclusion. Isolated VCF, especially those involving thoracic between vertebral compression fractures (VCFs) and pneumonia. region, was associated with increased risk of pneumonia. Summary of Background Data. VCF is known to be associ- Key words: compression fracture, osteopenia, osteoporosis, ated with impaired pulmonary function. However, the risk of pneumonia. developing pneumonia after new-onset VCF has not been Level of Evidence: 4 evaluated. Spine 2018;43:E830–E835 Methods. The data source was the National Health Screening Program Cohort, a population-based sample cohort of national health screening enrolees. Elderly women with low bone density identified during the life transition period health check-up were ow bone density due to osteopenia or osteoporosis is target population. Patients with newly diagnosed isolated VCF very common among elderly women.1 One of the after the health check-up except those with cervical VCF were most frequent complications of low bone density is included as exposure group. Matched individuals still free of any L 2,3 vertebral compression fractures (VCFs). In the United type of VCF from the same population were included as control States, approximately 1.5 million VCFs occur every year, group with case/control ratio 3:1. The time to first occurrence of and the estimated annual cost of VCFs in the US approaches pneumonia were analyzed with Cox-regression analysis. $746 million.4,5 Many studies exist in the literature about Results. From the target population (N ¼ 24,773), we matched the complications of VCFs, such as increased osteoporosis 867 patients exposed to new-onset isolated VCFs with 2601 due to inactivity, subsequent fractures, impaired pulmonary controls. In a multivariable Cox-regression analysis, isolated VCF function, and emotional problems.5–8 was identified as an independent risk factor for development of Pneumonia is one of the most common acute pulmo- a pneumonia (adjusted hazard ratio, 1.48; 95% confidence nary diseases that lead to a high mortality in elderly interval 1.14–1.91). Among the subtypes of VCF, thoracic/ patients.9 The incidence of hospitalization due to com- thoracolumbar VCF was independently associated with munity-acquired pneumonia approaches 0.6 million each year, and the incidence rises with increasing age.10,11 It is well known that there are many risk factors for pneumo- From the Department of Emergency Medicine, Seoul National University nia, from lifestyle factors such as smoking or alcohol Bundang Hospital, Gyeonggi-do, Republic of Korea; yDivision of Cardiol- abuse to underlying comorbidities such as chronic respi- ogy, Department of Internal Medicine, Seoul National University Bundang Hospital, Gyeonggi-do, Republic of Korea; and zDivision of Pulmonary and ratory and cardiovascular diseases, Parkinson disease, and 12,13 Critical Care Medicine, Department of Internal Medicine, Seoul National osteoporosis. University Bundang Hospital, Gyeonggi-do, Republic of Korea. Several studies have reportedanassociationbetween Acknowledgment date: September 27, 2017. First revision date: December thoracic injury such as rib fractures, and pneumonia.14–16 6, 2017. Acceptance date: December 8, 2017. It has been postulated that acute pain from such trauma The manuscript submitted does not contain information about medical device(s)/drug(s). could impair coughing and also secretion clearance, lead- 15 No funds were received in support of this work. ing to atelectasis and subsequent pneumonia. In VCF, No relevant financial activities outside the submitted work. spinal deformities such as thoracic kyphosis and back pain Address correspondence and reprint requests to Joonghee Kim, MD, MS, canleadtosimilarproblems.Onesystematicreview Department of Emergency Medicine, Seoul National University Bundang reported that, each single VCF leads to a 9% decrease Hospital, 166 Gumi-ro, Bundang-gu, Seongnam-si, Gyeonggi-do 463–707, in vital capacity.8 However, no study reporting whether Republic of Korea; E-mail: [email protected] acute VCF is associated with an increased risk of DOI: 10.1097/BRS.0000000000002536 pneumonia. E830 www.spinejournal.com July 2018 Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. EPIDEMIOLOGY Risk of Pneumonia After VCF in Women with Low Bone Density Kim et al

In this study, we postulated that VCF is associated with new VCF event to include only the patients with isolated an increased risk of pneumonia and conducted a population- vertebral fractures without significant accompanying injury. based observational study to prove the hypothesis. Three controls for each of the patients with the exposure who were still free of VCFs were randomly sampled from the METHODS same target population. The criterion ‘‘free of vertebral fracture’’ was defined as no appearance of the diagnostic Data Source codes for at least the same 5 years until a new vertebral The data source was the National Health Insurance Ser- fracture event was observed in the matched patient with vice—National Health Screening Program Cohort, a ran- exposure. If a patient who had a VCF at later time was dom sample cohort of national health screening enrollees sampled as a control for other case patient before its own recruited in 2002 and 2003 (10% of 5,150,000 enrollees controls were sampled, the patient was excluded from the ranged in age from 40 to 79 years in December 2003), exposure group. The main exposure variables were the released by the National Health Insurance Service in 2015 occurrence of a VCF and the region of the VCF, categorized to support academic researches. These data comprise as follows: (1) thoracic/thoracolumbar, (2) lumbar only, or detailed claim information as well as the results of various (3) unspecified and control. health screening programs conducted from 2002 to 2013. The variables used for the matching process included The data also include diagnostic codes, prescription, and age and BMD categorized as osteopenia (1.0 > T- procedure codes and related costs, as well as demographic score > 2.5) and osteoporosis (T-score 2.5). The pri- information such as age, sex, and socioeconomic status. mary outcome event was the first occurrence of pneumo- The diagnostic codes follow the 6th-revision of Korean nia or death due to pneumonia. The diagnostic codes used Classification of Diseases, which was developed based on to define a pneumonia event were J10.0x, J11.0x, J12.x, the 10th-revision of International Classification of Diseases J13.x, J14.x, J15.x, J16.x, J17.x, and J18.x. coding system. The data source also includes information about disability and death based on the national disability Statistical Analysis registration data and death certificates, respectively. Categorical variables were reported using frequencies and Detailed descriptions of the cohort data have been pub- proportions, whereas continuous variables were reported lished previously.17,18 The institutional review board of the using the mean and standard deviation or medians and study hospitals approved the analysis and waived the interquartile ranges, as appropriate. Student t test, Wil- requirement for informed consent. coxon rank-sum test, the chi-squared test, or Fisher exact test were performed, as appropriate, for comparisons Case Selection and Data Handling between the groups. The target population was female patients with low bone We used Cox regression to model the subsequent risk of mass density (BMD; due to osteoporosis or osteopenia, T developing pneumonia after each new VCF event. Assuming score <1) measured at a Life Transition Period Health the actual compression fracture event occurred within a check-up. The health screening program, begun in Korea in month before the first entry of related diagnostic codes, January 2007, is offered only twice during one’s lifetime, at we assessed the incidence of pneumonia from the time point the ages 40 and 66 years. The BMD is measured at the 1 month before the fracture diagnosis entry to 2 years after second screening (age 66), and therefore, all the BMD the diagnosis. The covariates used for adjustment were age, measurements used in this study were measured around sex, and comorbidities such as diabetes mellitus, hyperten- that age. sion, ischemic heart disease (IHD), stroke, heart failure, The exposure group comprised patients with a new event chronic renal failure, advanced liver disease, chronic of VCF after the health check-up without such an event obstructive pulmonary disorder (COPD), and malignancy. during the previous 5 years until the new fracture diagnosis. The presences of the comorbidities were based on the claim The new VCF event was defined as a new database entry data of the 2-year preobservation period. We first con- with the following discharge diagnosis codes: M48.5, structed two full multivariable models for pneumonia M48.50x, M48.54x, M48.55x, M48.56x, M48.57x, including all the covariates and each of the main exposure M48.59x, M80.08x, M80.18x, M80.88x, M80.98x, variables (VCF occurrence and fracture region, respec- M84.48x, S32, S320, S32.02, S32.020, S32.03, S32.030, tively). We subsequently constructed similar models using S32.04, S32.040, S32.05, S32.050, S32.06, S32.060, a variable selection scheme based on Akaike information S32.06, S32.060, S32.09, S32.090, S32.7, S32.70, S32.8, criterion. The goodness of fit of the models was tested with S32.83, S32.830, S22.0, S22.02, S22.020, S22.03, S22.030, Gronnesby and Borgan goodness-of-fit test, and their pro- S22.04, S22.040, S22.05, S22.050, S22.06, S22.060, portional hazard assumption was assessed by examination S22.07, S22.070, S22.09, S22.090, S22.1, and S22.10. of Schoenfeld residual plots. The results of the Cox regres- We only considered thoracic, thoracolumbar, or lumbar sion analyses were presented as hazard ratios (HRs) with fracture because cervical fractures are more likely to repre- their 95% confidence intervals (CIs). P values less than 0.05 sent high energy fractures. We excluded patients with any were considered significant. All the data handling and injury code (S- or T-code) other than those used to define the statistical analyses were performed using R-packages, Spine www.spinejournal.com E831 Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. EPIDEMIOLOGY Risk of Pneumonia After VCF in Women with Low Bone Density Kim et al

TABLE 1. Baseline Characteristics of the Study Populations Exposure Control (N ¼ 867) (N ¼ 2601) P Age, mean (SD) 68.6 (1.7) 68.6 (1.7) 1.000 Bone density 1.000 Osteopenia (%) 325 (37.5%) 975 (37.5%) Osteoporosis (%) 542 (62.5%) 1626 (62.5%) Region of involvement Thoracic/thoracolumbar (%) 253 (29.2%) Lumbar only (%) 401 (46.3%) N/A Unspecified (%) 213 (24.6%) Hypertension (%) 477 (55.0%) 1478 (56.8%) 0.374 Ischemic heart disease (%) 130 (15.0%) 327 (12.6%) 0.077 Heart failure (%) 53 (6.1%) 109 (4.2%) 0.026 Stroke (%) 72 (8.3%) 158 (6.1%) 0.027 Diabetes mellitus (%) 136 (15.7%) 423 (16.3%) 0.729 Chronic renal failure (%) 3 (0.3%) 24 (0.9%) 0.147 Advanced liver disease (%) 5 (0.6%) 15 (0.6%) 1.000 COPD (%) 33 (3.8%) 75 (2.9%) 0.214 Malignancy (%) 51 (5.9%) 155 (6.0%) 1.000 Incidence of pneumonia after VCF For 3 months (%) 41 (1.6%) 23 (2.7%) 0.058 For 2 years (%) 178 (6.8%) 87 (10.0%) 0.003 COPD indicates chornic obstructive pulmonary disease; SD, standard deviation; VCF, vertebral compression fracture.

version 3.3.2 (R Foundation for Statistical Computing, increased risk of pneumonia compared with the control Vienna, Austria). group (P < 0.001), based on the log-rank test. Table 2 shows the association between VCF and new- RESULTS onset pneumonia, based on univariable and full and reduced multivariable Cox regression models. With no adjustment, Baseline Characteristics VCF was associated with an increased risk of pneumonia, A total of 24,773 women at risk of VCF were identified from with an HR of 1.51 (95% CI 1.17–1.95). With adjustment, the health screening enrollees. From this population, a total VCF independently acted as a risk factor for pneumonia, 867 patients with VCF were included in the exposure group, both in the full model (HR ¼ 1.48, 95% CI 1.14–1.91) and and 2601 controls (a total three times the size of the the reduced model (HR ¼ 1.48, 95% CI 1.14–1.91). Other exposure group) were randomly sampled and included as independent risk factors included a history of IHD a control group (Table 1). The distributions of age and low (HR ¼ 1.66, 95% CI 1.20–2.30) and COPD (HR ¼ 2.66, BMD category were identical between the two groups as 95% CI 1.70–4.17) in the full model. The same risk factors, they were matched. The prevalence of comorbidities was not IHD (HR ¼ 1.64, 95% CI 1.21–2.22) and COPD significantly different between the groups except for heart (HR ¼ 2.72, 95% CI 1.74–4.25) were identified in the failure and stroke, which were both more frequent in the reduced model. exposure group. The most common region of VCF was the Table 3 shows the association between the anatomical lumbar area (46.3%), followed by thoracic/thoracolumbar regions of VCF and the occurrence of pneumonia. Only (29.2%) region and unspecified regions (24.6%). thoracic/thoracolumbar VCF was found to be an independent risk factor for pneumonia (univariable model: Impact of Vertebral Compression Fracture on HR ¼ 2.17, 95% CI 1.50–3.13; full model: HR ¼ 2.13, Pneumonia 95% CI 1.47–3.08; and reduced model: HR ¼ 2.11, 95% Figure 1A shows the cumulative incidence of pneumonia in CI 1.46–3.05). Other VCFs not involving the thoracic the VCF and control groups. The cumulative incidence of vertebrae did not represent significant risk factors pneumonia was significantly higher in the VCF group than for pneumonia. in the control group (P ¼ 0.002), based on the log-rank test. Figure 1B shows the cumulative incidence of pneumonia in DISCUSSION the VCF subtypes. Only the VCF subtype involving the In this study, isolated VCF was an independent risk factor thoracic/thoracolumbar region was associated with an for future development of pneumonia and, interestingly,

E832 www.spinejournal.com July 2018 Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. EPIDEMIOLOGY Risk of Pneumonia After VCF in Women with Low Bone Density Kim et al

Figure 1. A, Cumulative incidence of pneumonia after fracture events. B, Cumulative incidence of pneumonia by regions of involvement.

VCFs involving thoracic region was responsible for the rib fracture or other blunt trauma.15,20 However, the rather association. The results of this study support our hypothesis prolonged effect of VCF as shown in Figure 1 suggests that of the association between VCF and pneumonia. To our best the increased risk of pneumonia does not appear to be solely knowledge, this is the first study reporting the association dependent on acute pain, immobilization, or temporal brace between the two disease entities. application which could restrict vital capacity.21 It is possi- The association between VCF and pneumonia found in ble the decreased overall physical performance of the this study would have multiple explanations. First, it has patients with VCF may have prolonged effect such as been previously reported that VCFs are associated with increased nursing home dependency that might increase impaired pulmonary function including decrease in vital the chance of subsequent pneumonia development.22 capacity and inspiratory time.7,8,19 The restriction of active Clinically, it is interesting that VCFs involving thoracic cough due to pain and immobilization during the acute region was responsible for the association found in this phase may increase the risk of pneumonia, similar to minor study. In a previous study, thoracic spine injuries with

TABLE 2. Cox-Regression Hazard Models for the Presence of Vertebral Compression Fracture Model Variable Hazard Ratio (95% CI) P Univariable model Fracture 1.51 (1.17–1.95) 0.002 Multivariable model Fracture 1.48 (1.14–1.91) 0.003 Osteoporosis (relative to osteopenia) 1.2 (0.93–1.55) 0.168 Old age (70 or more) 0.96 (0.72–1.28) 0.783 Hypertension 0.88 (0.68–1.13) 0.313 Ischemic heart disease 1.66 (1.2–2.3) 0.002 Heart failure 1.38 (0.84–2.27) 0.210 Stroke 0.82 (0.5–1.35) 0.431 Diabetes mellitus 1.18 (0.86–1.63) 0.297 Chronic renal failure 1.2 (0.38–3.81) 0.759 Advanced liver disease 1.48 (0.37–6.01) 0.582 Chronic obstructive pulmonary disease 2.66 (1.7–4.17) 0.000 Malignancy 1.48 (0.95–2.29) 0.080 Reduced model (AIC criteria) Fracture 1.48 (1.14–1.91) 0.003 Ischemic heart disease 1.64 (1.21–2.22) 0.001 Chronic obstructive pulmonary disease 2.72 (1.74–4.25) 0.000 Malignancy 1.5 (0.97–2.32) 0.068 AIC indicates Akaike information criterion; CI, confidence interval.

Spine www.spinejournal.com E833 Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. EPIDEMIOLOGY Risk of Pneumonia After VCF in Women with Low Bone Density Kim et al

TABLE 3. Cox-Regression Risk Models for Vertebral Compression Fracture Regions Model Variable Hazard Ratio (95% CI) P Univariable model Thoracic/thoracolumbar 2.17 (1.5–3.13) 0.000 Lumbar only 1.2 (0.83–1.75) 0.338 Unspecified region 1.38 (0.88–2.17) 0.165 No fracture Reference Multivariable model Thoracic/thoracolumbar 2.13 (1.47–3.08) 0.000 Lumbar only 1.16 (0.8–1.69) 0.440 Unspecified region 1.36 (0.87–2.15) 0.181 No fracture Reference Osteoporosis (relative to osteopenia) 1.18 (0.91–1.53) 0.201 Old age (70 or more) 0.95 (0.71–1.26) 0.704 Hypertension 0.87 (0.67–1.13) 0.309 Ischemic heart disease 1.66 (1.2–2.29) 0.002 Heart failure 1.37 (0.83–2.26) 0.219 Stroke 0.82 (0.5–1.36) 0.445 Diabetes mellitus 1.21 (0.88–1.67) 0.235 Chronic renal failure 1.11 (0.35–3.53) 0.862 Advanced liver disease 1.59 (0.39–6.45) 0.519 Chronic obstructive pulmonary disease 2.68 (1.71–4.2) 0.000 Malignancy 1.46 (0.94–2.27) 0.089 Reduced model (AIC criteria) Thoracic/thoracolumbar 2.11 (1.46–3.05) 0.000 Lumbar only 1.15 (0.79–1.68) 0.461 Unspecified region 1.4 (0.89–2.2) 0.148 No fracture Reference Ischemic heart disease 1.64 (1.21–2.22) 0.001 Chronic obstructive pulmonary disease 2.74 (1.75–4.29) 0.000 Malignancy 1.49 (0.96–2.3) 0.076 AIC indicates Akaike information criterion; CI, confidence interval. accompanying spine cord injury (SCI) were associated with particular, we could not distinguish the types of pneumonia increased respiratory complications because of the increased that occur predominantly after VCF. Second, we only con- respiratory muscle paralysis or interruption of sympathetic sidered the first events of VCF. Considering isolated VCF airway innervation.16 Because approximately 10% of SCIs would serves as a risk factor for a subsequent VCF that are caused by mild compression fractures, SCIs resulting might cause impaired pulmonary function, it is unclear from thoracic VCFs may play a role in increase the risk of whether the development of pneumonia was affected by a pneumonia.23 In this study, we only included isolated VCF single VCF event or repeated VCFs. Finally, the data source while excluding those with diagnostic codes for accompa- does not provide any detailed information of the VCF nying injuries such as those for traumatic CNS injuries. incidences such as injury severity, mechanisms, or other Comorbidities such as COPD and IHD are well known accompanying injuries such as SCI. Although we excluded risk factors for pneumonia.12,24,25 COPD and osteoporosis cases with any diagnostic codes for accompanying injuries are strongly associated because they share several risk fac- to consider only isolated VCFs, detailed clinical information tors such as age, immobility, smoking, and the use of describing the incidences would be helpful in improving the corticosteroids.26,27 In addition, there have been many accuracy of initial case selection. studies that report acute exacerbations of COPD patients Despite those limitations, this study has several strengths. after VCFs.28,29 Consistent with previous studies, both Because this is a population-based study, we could generate COPD and IHD were independent risk factors of pneumo- randomly sampled controls who had not exposed to VCF nia in this study. Therefore, it would be considerate to pay from the population at risk for VCF. If this was hospital close attention to the pulmonary function of the patients registry–based study, such identification of control group with VCF if a patient has such comorbidities. Recommend- would be difficult. Secondly, relatively long-term observa- ing incentive spirometry or deep breathing exercises might tion could be possible. We think long-term observation is be beneficial in such cases.30 important because lasting deterioration of pulmonary func- This study has several limitations. First, the study is based tion or performance due to VCF might play a significant role on claim data in which the accuracy of diagnostic codes is in increasing the risk of pneumonia after VCF. Finally risk limited and significant over- or undercoding could exist. In stratification based on the region of VCF involvement may

E834 www.spinejournal.com July 2018 Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. EPIDEMIOLOGY Risk of Pneumonia After VCF in Women with Low Bone Density Kim et al

be of useful in predicting future risk of pneumonia and 9. Simonetti AF, Viasus D, Garcia-Vidal C, et al. Management of community-acquired pneumonia in older adults. Ther Adv Infect planning future treatment. Dis 2014;2:3–16. In conclusion, isolated VCF was associated with 10. Kaplan V, Angus DC, Griffin MF, et al. Hospitalized community- increased risk of pneumonia. Physicians and caregivers acquired pneumonia in the elderly: age- and sex-related patterns of taking care of VCF patients should be vigilant for develop- care and outcome in the United States. Am J Respir Crit Care Med 2002;165:766–72. ment of pneumonia especially if the VCF involves thoracic 11. Majumdar SR, McAlister FA, Eurich DT, et al. Statins and out- region. comes in patients admitted to hospital with community acquired pneumonia: population based prospective cohort study. BMJ 2006;333:999. Key Points 12. Torres A, Peetermans WE, Viegi G, et al. Risk factors for community-acquired pneumonia in adults in Europe: a literature review. Pulmonary function deterioration after VCFs has Thorax 2013;68:1057–65. 13. Vinogradova Y, Hippisley-Cox J, Coupland C. Identification of been known, but the occurrence of pneumonia new risk factors for pneumonia: population-based case-control after the compression fractures has not been study. Br J Gen Pract 2009;59:e329–38. studied. 14. Brasel KJ, Guse CE, Layde P, et al. Rib fractures: relationship with Using National Health Insurance Service-National pneumonia and mortality. Crit Care Med 2006;34:1642–6. 15. Ho SW, Teng YH, Yang SF, et al. Risk of pneumonia in patients Sample Cohort representing 2.2% of Korean with isolated minor rib fractures: a nationwide cohort study. BMJ population, we found that an adjusted HR for Open 2017;7:e013029. pneumonia in the women with VCF was 1.48 16. Cotton BA, Pryor JP, Chinwalla I, et al. Respiratory complications compared with control group. and mortality risk associated with thoracic spine injury. J Trauma Especially VCFs in only thorax region represented 2005;59:1400–7; discussion 1407–1409. 17. Lee YS, Lee YR, Chae Y, et al. Translation of Korean medicine use as an independent risk factor for pneumonia to ICD-codes using National Health Insurance Service-National compare with other regions. Sample Cohort. Evid Based Complement Alternat Med 2016; In the context of increased burden of disease 2016:8160838. caused by osteoporotic compression fractures, 18. Kim B, Jeong H, Kim J, et al. Incidence and risk factors of delayed intracranial hemorrhage in emergency department. Am J Emerg physicians should know that the compression Med 2018;36:271–6. fractures can act as a long-term risk factor for 19. Yang HL, Zhao L, Liu J, et al. Changes of pulmonary function for pneumonia. patients with osteoporotic vertebral compression fractures after kyphoplasty. J Spinal Disord Tech 2007;20:221–5. 20. Kea B, Gamarallage R, Vairamuthu H, et al. What is the clinical significance of chest CT when the chest x-ray result is normal in patients with blunt trauma? Am J Emerg Med 2013;31:1268–73. Acknowledgments 21. Voormolen MH, Mali WP, Lohle PN, et al. Percutaneous verte- The authors thank to the Division of Statistics in Medical broplasty compared with optimal pain medication treatment: Research Collaborating Center at Seoul National University short-term clinical outcome of patients with subacute or chronic painful osteoporotic vertebral compression fractures. The VER- Bundang Hospital for statistical analysis. TOS study. AJNR Am J Neuroradiol 2007;28:555–60. 22. Muder RR. Pneumonia in residents of long-term care facilities: epidemiology, etiology, management, and prevention. Am J Med References 1998;105:319–30. 1. Reginster JY, Burlet N. Osteoporosis: a still increasing prevalence. 23. Sekhon LH, Fehlings MG. Epidemiology, demographics, and Bone 2006;38 (2 suppl 1):S4–9. pathophysiology of acute spinal cord injury. Spine (Phila Pa 2. Johnell O, Kanis JA. An estimate of the worldwide prevalence and 1976) 2001;26 (24 suppl):S2–12. disability associated with osteoporotic fractures. Osteoporos Int 24. Mannino DM, Davis KJ, Kiri VA. Chronic obstructive pulmonary 2006;17:1726–33. disease and hospitalizations for pneumonia in a US cohort. Respir 3. Papaioannou A, Watts NB, Kendler DL, et al. Diagnosis and Med 2009;103:224–9. management of vertebral fractures in elderly adults. Am J Med 25. Lin SH, Perng DW, Chen CP, et al. Increased risk of community- 2002;113:220–8. acquired pneumonia in COPD patients with comorbid cardiovas- 4. Lad SP, Patil CG, Lad EM, et al. Trends in pathological vertebral cular disease. Int J Chron Obstruct Pulmon Dis 2016;11:3051–8. fractures in the United States: 1993 to 2004. J Neurosurg Spine 26. Lehouck A, Boonen S, Decramer M, et al. COPD, bone metabo- 2007;7:305–10. lism, and osteoporosis. Chest 2011;139:648–57. 5. Alexandru D, So W. Evaluation and management of vertebral 27. Sarkar M, Bhardwaj R, Madabhavi I, et al. Osteoporosis in chronic compression fractures. Perm J 2012;16:46–51. obstructive pulmonary disease. Clin Med Insights Circ Respir 6. Lindsay R, Burge RT, Strauss DM. One year outcomes and costs Pulm Med 2015;9:5–21. following a vertebral fracture. Osteoporos Int 2005;16:78–85. 28. Papaioannou A, Parkinson W, Ferko N, et al. Prevalence of 7. Krege JH, Kendler D, Krohn K, et al. Relationship between vertebral fractures among patients with chronic obstructive pul- vertebral fracture burden, height loss, and pulmonary function monary disease in Canada. Osteoporos Int 2003;14:913–7. in postmenopausal women with osteoporosis. J Clin Densitom 29. Nuti R, Siviero P, Maggi S, et al. Vertebral fractures in patients 2015;18:506–11. with chronic obstructive pulmonary disease: the EOLO Study. 8. Harrison RA, Siminoski K, Vethanayagam D, et al. Osteoporosis- Osteoporos Int 2009;20:989–98. related kyphosis and impairments in pulmonary function: a sys- 30. Kelkar KV. Post-operative pulmonary complications after non- tematic review. J Bone Miner Res 2007;22:447–57. cardiothoracic surgery. Indian J Anaesth 2015;59:599–605.