Individuals with ESRD experience increased cardiovascular morbidity and mortality compared with the general population (1). ESRD patients are at high risk for vascular calcification, a highly regulated process that appears to be partially mediated by disorders of mineral metabolism (2). In response to various stimuli, vascular smooth muscle cells undergo a phenotypic change characterized by expression of proteins that promote bone formation, including alkaline phosphatase (AlkPhos) (3). Phosphate induces this phenotypic change in vitro (2) and upregulates AlkPhos activity (4). Thus AlkPhos and phosphate have been directly implicated in the pathogenesis of vascular calcification and subsequent cardiovascular disease (CVD).
In patients with ESRD and in the general population, higher serum levels of AlkPhos and phosphate are associated with increased all-cause and cardiovascular mortality (5–13). Recent studies support the hypothesis that these associations are mediated by vascular calcification (14,15), although medial calcification is seen almost exclusively in patients with advanced kidney disease and diabetes and not in the general population (16). These findings are clinically relevant because medial and intimal vascular calcification have been shown to predict cardiovascular risk in dialysis patients (17) and the general population (18). In addition, several studies have demonstrated an association between vascular calcification and low bone density (19–21), suggesting a role for bone mineral parameters as a marker of both processes.
On the basis of these data, we evaluated the association between AlkPhos and phosphate and subsequent mortality and hospitalizations in patients with an estimated GFR (eGFR) ≥60 ml/min/1.73 m2 in a large medical system in the Bronx, New York.
Materials and Methods
Study Population
Montefiore Medical Center (MMC) is a large tertiary-care center in the Bronx, New York. Individuals were eligible for inclusion if they were seen as outpatients at MMC between January 1, 2000 and December 31, 2002; were 18 years of age or older at the time of the index visit (n = 205,294); and had levels of serum AlkPhos, calcium, phosphate, albumin, and creatinine checked as outpatients within 2 days of the visit (n = 20,573). If an individual had multiple visits during this time period, the first was considered the index visit. Patients were excluded if they had an eGFR <60 ml/min/1.73 m2 or had ever received treatment with any dialysis modality (n = 3623); or had a history of cancer, liver or biliary disease, HIV/AIDS, or organ transplantation on the basis of inpatient or outpatient International Classification of Diseases, Ninth Revision (ICD-9) codes or subspecialty clinic visits (n = 6207). A total of 10,743 individuals were included in the final analyses. The Committee on Clinical Investigation at the Albert Einstein College of Medicine and MMC approved the study protocol.
Data Collection
Data were extracted using the Clinical Looking Glass system (Emerging Health Information Technology, Yonkers, NY), an interactive software application developed at MMC that integrates clinical and administrative data sets (22). All laboratory values aside from those required for inclusion in the study were measured within 7 days of the baseline AlkPhos and phosphate measurements. ICD-9 codes before the index date were used to define baseline hypertension (401 to 405, 437.2) and CVD (410 to 414, 427 to 440). Baseline diabetes mellitus (DM) was defined as an ICD-9 code for DM (250, 357.2), serum glucose values ≥200 mg/dl, or a hemoglobin A1c ≥7% on two occasions before the index date.
Serum chemistry values were measured using the Hitachi Modular System (Roche Diagnostics, Indianapolis, IN). Serum AlkPhos (normal range, 30 to 115 U/L) was measured using the cleavage of p-nitrophenyl phosphate to p-nitrophenol and phosphate, in which p-nitrophenol release is proportional to AlkPhos activity and is measured photometrically. Serum phosphate (normal range, 2.5 to 4.5 mg/dl) was measured using an ammonium molybdate assay. Serum creatinine was measured by a modified kinetic Jaffé reaction via the phosphoenolpyruvate carboxylase method. eGFR was calculated using the four-variable Modification of Diet in Renal Disease study equation (23). eGFR values were truncated at 200 ml/min/1.73 m2 (n = 61) because levels above this value were thought to be physiologically implausible. Serum calcium values were corrected for serum albumin concentration using the formula corrected total calcium (mg/dl) = total calcium (mg/dl) + 0.8 [4 − albumin (g/dl)].
Outcome Variables
Data regarding follow-up were available through September 11, 2008. The main outcome variables were all-cause mortality and hospitalizations. Clinical Looking Glass merges its data monthly with the Social Security Death Registry, a validated method for the assessment of mortality (24). Cause-specific hospitalizations were defined as admissions to MMC with ICD-9 codes at the time of discharge for CVD (410 to 414, 427 to 440), fracture (800 to 829), or infection (001 to 139, 320 to 324, 421, 460 to 466, 480 to 486, 590, 595, 599, 680 to 686, 790.7, 790.8), respectively. As many of the baseline laboratory tests may have been obtained in the setting of preprocedural or presurgical evaluation, we excluded hospitalizations that occurred within 28 days of the baseline outpatient visit.
Statistical Analyses
Characteristics of the population categorized by quartiles of serum AlkPhos and serum phosphate, separately, were compared using χ2 for categorical variables and analysis-of-variance or the Kruskal–Wallis test for continuous variables. Cox proportional hazards models were created to separately evaluate the risk of mortality and hospitalization. Age; sex; race/ethnicity; diagnosis of DM, hypertension, and CVD; eGFR; and serum albumin were included a priori in all models, as well as additional covariates associated with serum AlkPhos or phosphate or the outcome at P < 0.20. To account for possible confounding by indication related to preprocedural testing, we adjusted for hospitalization within 28 days after the index date. The final models also included insurance status, serum AlkPhos, phosphate, corrected total calcium, bicarbonate, cholesterol, serum glutamic pyruvic transaminase (SGPT), total bilirubin, and hemoglobin. Additional models were examined including serum AlkPhos or phosphate, separately. Models were also created including all potential covariates regardless of associations with serum AlkPhos, phosphate, or the outcome. All analyses were repeated using eGFR calculated with the Chronic Kidney Disease Epidemiolgy Collaboration equation (25). Because results for each outcome were not different from the final models, only the results of the final models are presented. To test effect modification between serum AlkPhos and phosphate, an interaction term was included in the model. Individuals were further categorized by levels of serum AlkPhos and phosphate above or below the median for each. The hazard ratio (HR) for mortality associated with high phosphate alone, high AlkPhos alone, and both of these factors versus individuals with neither factor was calculated. Because of the small number of events (n = 144), hospitalization related to fracture was examined within tertiles of serum AlkPhos and phosphate. For the hospitalization outcomes, patients were censored at the time of death or, if they remained alive, at the time of their last evaluation in the MMC system. The proportionality assumption was shown to be accurate by visual inspection of log-log plots. Missing values for covariates were imputed using multiple imputation methods and including among the predictors follow-up survival status and the log of the survival time (26,27). Twelve percent of serum cholesterol values and ≤1% of the values for other covariates used in the multivariable analyses were missing and thus imputed.
Sensitivity Analyses
To evaluate the effect of residual confounding because of liver or biliary disease on the association of AlkPhos with mortality and hospitalizations, we excluded all patients with a history of serum glutamic oxaloacetic transaminase >40 U/L, SGPT >40 U/L, gamma glutamyltransferase >54 U/L, positive antibody test to hepatitis C, or positive serology for hepatitis B surface antigen at any time before the index date. We then further excluded patients with baseline AlkPhos >115 U/L (the upper limit of normal in our laboratory). Because AlkPhos and phosphate testing might have been prompted by acute and subacute illness, we repeated the analysis of the mortality end point with a 2-year time lag after the index date, and also after excluding patients who were hospitalized in the first 28 days after the index date, using the baseline values of serum AlkPhos and phosphate.
Analyses of hospitalization outcomes were repeated using a 6-month time lag after baseline AlkPhos and phosphate measurements. In addition, because individuals residing near MMC may be more likely to be hospitalized there than residents further removed geographically, we limited our analysis of hospitalization outcomes to those residing in the Bronx at the time of the index visit. We also performed a separate analysis among those who had an inpatient or outpatient encounter in the MMC system >6 months before the index date to allow time for the diagnosis of comorbidities. Statistical analyses were performed using Stata software, version 10.0 (Stata Corporation, College Station, TX). P < 0.05 was considered statistically significant.
Results
Patient Characteristics
The mean age of the 10,743 patients was 51 years, 64% were women, 22% were white, 26% were non-Hispanic black, 16% were Hispanic, 13% had a diagnosis of hypertension, 9% had a diagnosis of DM, and 8% had prior CVD. Baseline characteristics of the patients by quartiles of serum AlkPhos and phosphate are shown in Tables 1 and 2, respectively. Compared with those excluded (n = 194,551), included patients were older and more likely to be women and to be white (P ≤ 0.001 for each).
;)
Table 1: Baseline characteristics of 10,743 outpatients by quartiles of serum AlkPhosa,b,c
Table 2: Baseline characteristics of 10,743 outpatients by quartiles of serum phosphatea,b
Associations between Serum AlkPhos and Phosphate and Mortality
The median follow-up time was 6.8 years (interquartile range 6.1 to 7.6). There were 949 patients who died during follow-up. After multivariable adjustment, higher values of serum AlkPhos and phosphate were associated with an increased risk of death (Table 3). Compared with patients with levels of serum AlkPhos and phosphate below the median for each (<83 U/L and <3.4 mg/dl, respectively), those with only phosphate above the median did not have a significantly increased risk of death after multivariable adjustment [HR 1.15, 95% confidence interval (CI) 0.93 to 1.41]. Those with AlkPhos but not phosphate above the median or both above the median had an increased risk of death (adjusted HR 1.42, 95% CI 1.18 to 1.71; and adjusted HR 1.65, 95% CI 1.37 to 1.98, respectively). Hospitalization within 28 days after baseline laboratory testing was not associated with mortality (P = 0.45).
Table 3: Unadjusted and adjusted HR of all-cause mortality by baseline serum AlkPhos and phosphate level in 10,743 outpatientsa
Associations between Serum AlkPhos and Phosphate and Hospitalizations
Overall, 4237 patients were hospitalized during the follow-up period. There were 1662 cardiovascular-related hospitalizations, 1142 infection-related hospitalizations, and 144 fracture-related hospitalizations. After multivariable adjustment, higher values of serum AlkPhos were associated with an increased risk of hospitalization because of any cause, CVD, infection, or fracture, and higher values of serum phosphate were associated with an increased risk of CVD-related hospitalization (Table 4). Hospitalization within 28 days after baseline laboratory testing was not associated with subsequent hospitalization overall (P = 0.89).
Table 4: Adjusted HR of hospitalization by baseline serum AlkPhos and phosphate level in 10,743 outpatientsa,b
Sensitivity Analyses
Mortality.
Multiple sensitivity analyses did not alter the association of serum AlkPhos with mortality, and there remained a trend toward increased risk of mortality with higher serum phosphate (Table 5). Furthermore, excluding patients with any history of a liver test abnormality or baseline serum AlkPhos >115 U/L (n included in analysis = 7685), the multivariable adjusted HR for the highest quartile (92 to 115 U/L) was 1.37 (95% CI 1.09 to 1.73) compared with the lowest quartile (≤63 U/L). Serum AlkPhos, modeled as a continuous variable, remained a significant predictor of mortality among this subgroup of individuals with serum AlkPhos ≤115 U/L [HR 1.38 per SD (45 U/L) increase, 95% CI 1.13 to 1.69].
Table 5: Sensitivity analyses examining adjusted HR of all-cause mortality among specific subpopulations by baseline serum AlkPhos and phosphate levela,b
Hospitalizations.
In sensitivity analyses, all hospitalization outcomes were examined after excluding those which occurred in the first 6 months after the baseline laboratory measures. There remained an increased risk of hospitalization for patients in the highest quartile of AlkPhos compared with those in the lowest: all-cause hospitalization, HR 1.20 (95% CI 1.07 to 1.34); CVD hospitalization, HR 1.25 (95% CI 1.05 to 1.47); and infection hospitalization, HR 1.42 (95% CI 1.16 to 1.73). The HR for fracture hospitalization in the highest versus lowest tertile of AlkPhos was 1.60 (95% CI 1.01 to 2.56). For those in the highest quartile of phosphate compared with the lowest, the HR for CVD-related hospitalization was 1.17 (95% CI 1.00 to 1.36). Restricting the analysis to patients residing in the Bronx or to those in the MMC system for at least 6 months also did not change the associations (data not shown).
Discussion
Higher serum concentrations of AlkPhos and phosphate in our analysis, even within the normal range, were associated with an increased risk of all-cause mortality and cardiovascular-related hospitalization in individuals with eGFRs ≥60 ml/min/1.73 m2. Higher AlkPhos levels were also associated with greater risk of hospitalization due to any cause or related to infection or fracture. These results remained robust after multiple sensitivity analyses. Our findings extend previous associations of serum AlkPhos and phosphate with adverse outcomes in the general population (9,10,13) by examining hospitalization outcomes and they provide further support to the notion of a bone-vascular axis linking vascular calcification and bone health.
A growing body of evidence suggests a link between vascular disease and bone loss. Novel inhibitors of AlkPhos inhibit vascular smooth muscle cell calcification in vitro (28). Greater severity of aortic calcification has been associated with higher levels of bone-specific AlkPhos (29) and increased fracture risk in postmenopausal women (19,21). Given the lack of data on vascular calcification and bone density in our cohort, we may only speculate regarding serum AlkPhos as a marker of vascular and bone pathology.
The association between AlkPhos levels and infection in our cohort follows recent associations in dialysis patients of higher serum AlkPhos (12) and phosphate (30) levels with the risk of infection. Factors regulating osteoblast and osteoclast development influence B cell maturation (31). Thus our findings may reflect interplay between bone metabolism and the immune system.
Values of bone-specific AlkPhos would have enabled us to test the hypothesis that higher AlkPhos levels are simply a marker for subclinical hepatic steatosis and associated metabolic abnormalities, but they were unavailable. We examined associations with total AlkPhos because this test is routinely measured in clinical practice. Our study is strengthened by the exclusion of individuals with a history of liver disease, adjustment for SGPT and total bilirubin, and by sensitivity analyses excluding patients with a history of viral hepatitis or any previously abnormal liver test. Further limiting the analyses to patients with AlkPhos levels within the normal range did not change our results. Thus our findings are unlikely to be limited by significant residual confounding by liver disease.
Our study has several important limitations. Data were obtained from an administrative database and were not collected for research purposes. Therefore, information regarding comorbidities may be incomplete, and we did not have data about disorders of bone metabolism such as Paget's disease. The use of administrative claims data to define hospitalization outcomes has recognized limitations, including misclassification of outcomes (32), but has been used previously for hospitalizations related to CVD (33), infection (34), and fractures (35–37). The reasons for performing laboratory testing in each patient are not known; preprocedural evaluation may have prompted some testing. Sensitivity analyses addressing confounding by indication for laboratory testing did not substantively alter our results. As in any observational study, there could be confounding by unmeasured covariates such as vitamin D or parathyroid hormone. In addition, causality cannot be established because of the observational study design. Ascertainment of hospitalization was limited to follow-up at MMC. However, our results remained robust after restricting hospitalization analyses to patients residing in the Bronx. Because mortality was ascertained via the Social Security Death Registry, follow-up at MMC was not required for assessment of death. Finally, because the study was conducted among individuals seeking care at a tertiary-care inner-city medical center who had measurements of serum AlkPhos and phosphate, the findings may not be applicable to the general population.
In summary, higher serum AlkPhos and phosphate levels, even within the normal range, are associated with increased mortality and CVD-related hospitalization in a multi-ethnic inner-city cohort with eGFR ≥60 ml/min/1.73 m2. Higher AlkPhos levels also predicted an increased risk of hospitalization related to infection or fracture. These findings may be representative of a common pathophysiology linking bone and vascular calcification and, perhaps, immune function. Further studies are needed to better elucidate these mechanisms, including the role of AlkPhos inhibition in the prevention of human disease.
Disclosures
Dr. Hostetter has consulted for Bristol Myers Squibb, Eli Lilly, and Wyeth. None of the other authors have any financial conflicts to disclose.
This research was supported by National Institutes of Health (NIH) grants K23 DK078774 to Dr. Melamed and R21 DK 077326 and RO1 DK080123 to Dr. Hostetter; and Clinical and Transitional Science Awards grants UL1 RR025750, KL2 RR025749, and TL1 RR025748 from the National Center for Research Resources, a component of NIH. Portions of this work were presented at the National Kidney Foundation Spring Clinical Meeting in Dallas, Texas, April 2-6, 2008 and at the American Society of Nephrology Renal Week in San Diego, CA, October 30, 2009.
Published online ahead of print. Publication date available at www.cjasn.org.
References
1. Foley RN, Parfrey PS, Sarnak MJ: Clinical epidemiology of cardiovascular disease in chronic renal disease. Am J Kidney Dis 32: S112–S119, 1998
2. Giachelli CM, Speer MY, Li X, Rajachar RM, Yang H: Regulation of vascular calcification: Roles of phosphate and osteopontin. Circ Res 96: 717–722, 2005
3. Steitz SA, Speer MY, Curinga G, Yang HY, Haynes P, Aebersold R, Schinke T, Karsenty G, Giachelli CM: Smooth muscle cell phenotypic transition associated with calcification: Upregulation of Cbfa1 and downregulation of smooth muscle lineage markers. Circ Res 89: 1147–1154, 2001
4. Lomashvili K, Garg P, O'Neill WC: Chemical and hormonal determinants of vascular calcification in vitro. Kidney Int 69: 1464–1470, 2006
5. Block GA, Hulbert-Shearon TE, Levin NW, Port FK: Association of serum phosphorus and calcium x phosphate product with mortality risk in chronic hemodialysis patients: A national study. Am J Kidney Dis 31: 607–617, 1998
6. Block GA, Klassen PS, Lazarus JM, Ofsthun N, Lowrie EG, Chertow GM: Mineral metabolism, mortality, and morbidity in maintenance hemodialysis. J Am Soc Nephrol 15: 2208–2218, 2004
7. Ganesh SK, Stack AG, Levin NW, Hulbert-Shearon T, Port FK: Association of elevated serum PO(4), Ca × PO(4) product, and parathyroid hormone with cardiac mortality risk in chronic hemodialysis patients. J Am Soc Nephrol 12: 2131–2138, 2001
8. Melamed ML, Eustace JA, Plantinga L, Jaar BG, Fink NE, Coresh J, Klag MJ, Powe NR: Changes in serum calcium, phosphate, and PTH and the risk of death in incident dialysis patients: A longitudinal study. Kidney Int 70: 351–357, 2006
9. Dhingra R, Sullivan LM, Fox CS, Wang TJ, D'Agostino RB Sr, Gaziano JM, Vasan RS: Relations of serum phosphorus and calcium levels to the incidence of cardiovascular disease in the community. Arch Intern Med 167: 879–885, 2007
10. Tonelli M, Sacks F, Pfeffer M, Gao Z, Curhan G: Relation between serum phosphate level and cardiovascular event rate in people with coronary disease. Circulation 112: 2627–2633, 2005
11. Regidor DL, Kovesdy CP, Mehrotra R, Rambod M, Jing J, McAllister CJ, Van Wyck D, Kopple JD, Kalantar-Zadeh K: Serum alkaline phosphatase predicts mortality among maintenance hemodialysis patients. J Am Soc Nephrol 19: 2193–2203, 2008
12. Blayney MJ, Pisoni RL, Bragg-Gresham JL, Bommer J, Piera L, Saito A, Akiba T, Keen ML, Young EW, Port FK: High alkaline phosphatase levels in hemodialysis patients are associated with higher risk of hospitalization and death. Kidney Int 74: 655–663, 2008
13. Tonelli M, Curhan G, Pfeffer M, Sacks F, Thadhani R, Melamed ML, Wiebe N, Muntner P: Relation between alkaline phosphatase, serum phosphate, and all-cause or cardiovascular mortality. Circulation 120: 1784–1792, 2009
14. Foley RN, Collins AJ, Herzog CA, Ishani A, Kalra PA: Serum phosphorus levels associate with coronary atherosclerosis in young adults. J Am Soc Nephrol 20: 397–404, 2009
15. Shantouf R, Kovesdy CP, Kim Y, Ahmadi N, Luna A, Luna C, Rambod M, Nissenson AR, Budoff MJ, Kalantar-Zadeh K: Association of serum alkaline phosphatase with coronary artery calcification in maintenance hemodialysis patients. Clin J Am Soc Nephrol 4: 1106–1114, 2009
16. O'Neill WC, Sigrist MK, McIntyre CW: Plasma pyrophosphate and vascular calcification in chronic kidney disease. Nephrol Dial Transplant 25: 187–191, 2010
17. Blacher J, Guerin AP, Pannier B, Marchais SJ, London GM: Arterial calcifications, arterial stiffness, and cardiovascular risk in end-stage renal disease. Hypertension 38: 938–942, 2001
18. Detrano R, Guerci AD, Carr JJ, Bild DE, Burke G, Folsom AR, Liu K, Shea S, Szklo M, Bluemke DA, O'Leary DH, Tracy R, Watson K, Wong ND, Kronmal RA: Coronary calcium as a predictor of coronary events in four racial or ethnic groups. N Engl J Med 358: 1336–1345, 2008
19. Bagger YZ, Tanko LB, Alexandersen P, Qin G, Christiansen C: Radiographic measure of aorta calcification is a site-specific predictor of bone loss and fracture risk at the hip. J Intern Med 259: 598–605, 2006
20. Hak AE, Pols HA, van Hemert AM, Hofman A, Witteman JC: Progression of aortic calcification is associated with metacarpal bone loss during menopause: A population-based longitudinal study. Arterioscler Thromb Vasc Biol 20: 1926–1931, 2000
21. Schulz E, Arfai K, Liu X, Sayre J, Gilsanz V: Aortic calcification and the risk of osteoporosis and fractures. J Clin Endocrinol Metab 89: 4246–4253, 2004
22. Southern WN, Berger MA, Bellin EY, Hailpern SM, Arnsten JH: Hospitalist care and length of stay in patients requiring complex discharge planning and close clinical monitoring. Arch Intern Med 167: 1869–1874, 2007
23. Levey AS, Greene T, Kusek JW, Beck GJ: A simplified equation to predict glomerular filtration rate from serum creatinine [Abstract]. J Am Soc Nephrol 11: 155A, 2000
24. Cowper DC, Kubal JD, Maynard C, Hynes DM: A primer and comparative review of major US mortality databases. Ann Epidemiol 12: 462–468, 2002
25. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF III, Feldman HI, Kusek JW, Eggers P, Van Lente F, Greene T, Coresh J: A new equation to estimate glomerular filtration rate. Ann Intern Med 150: 604–612, 2009
26. Royston P: Multiple imputation of missing values. Stata J 4: 227–241, 2004
27. van Buuren S, Boshuizen HC, Knook DL: Multiple imputation of missing blood pressure covariates in survival analysis. Stat Med 18: 681–694, 1999
28. Narisawa S, Harmey D, Yadav MC, O'Neill WC, Hoylaerts MF, Millan JL: Novel inhibitors of alkaline phosphatase suppress vascular smooth muscle cell calcification. J Bone Miner Res 22: 1700–1710, 2007
29. Iba K, Takada J, Yamashita T: The serum level of bone-specific alkaline phosphatase activity is associated with aortic calcification in osteoporosis patients. J Bone Miner Metab 22: 594–596, 2004
30. Plantinga LC, Fink NE, Melamed ML, Briggs WA, Powe NR, Jaar BG: Serum phosphate levels and risk of infection in incident dialysis patients. Clin J Am Soc Nephrol 3: 1398–1406, 2008
31. Horowitz MC, Bothwell AL, Hesslein DG, Pflugh DL, Schatz DG: B cells and osteoblast and osteoclast development. Immunol Rev 208: 141–153, 2005
32. Jollis JG, Ancukiewicz M, DeLong ER, Pryor DB, Muhlbaier LH, Mark DB: Discordance of databases designed for claims payment
versus clinical information systems. Implications for outcomes research. Ann Intern Med 119: 844–850, 1993
33. Dominici F, Peng RD, Bell ML, Pham L, McDermott A, Zeger SL, Samet JM: Fine particulate air pollution and hospital admission for cardiovascular and respiratory diseases. JAMA 295: 1127–1134, 2006
34. Chavers BM, Solid CA, Gilbertson DT, Collins AJ: Infection-related hospitalization rates in pediatric
versus adult patients with end-stage renal disease in the United States. J Am Soc Nephrol 18: 952–959, 2007
35. Baxter NN, Habermann EB, Tepper JE, Durham SB, Virnig BA: Risk of pelvic fractures in older women following pelvic irradiation. JAMA 294: 2587–2593, 2005
36. Brauer CA, Coca-Perraillon M, Cutler DM, Rosen AB: Incidence and mortality of hip fractures in the United States. JAMA 302: 1573–1579, 2009
37. Parikh S, Mogun H, Avorn J, Solomon DH: Osteoporosis medication use in nursing home patients with fractures in 1 US state. Arch Intern Med 168: 1111–1115, 2008
