Excelsior College PBH 321 CRITICAL EYE ON RESEARCH IN EPIDEMIOLOGY By the end of this activity, you will be able to illustrate the purposes, designs, weaknesses, and relevance of a randomized drug trial and of a crosssectional study. In this module, you will learn to analyze a research article on an epidemiological topic. It is understood that you may have minimal background in some of the medical terminology and statistical methods described in the paper. However, comprehending these components fully is not critical to evaluating the experimental study design. After reading the article, you should attempt to answer the questions posed for this research. Once you have answered these questions to the best of your ability, you can read commentary provided by your instructor so that you can see how critical analysis is applied to a research article. In the graded portion of this activity, you will apply this same strategy to another article, answering questions with a critical eye on the research. The instructions below will help you complete this activity. Step 1 Read the article, Lane et al., Tanezumab for the Treatment of Pain from Osteoarthritis of the Knee. New England Journal of Medicine, 2010; 363; 16. You will find a link for this article in the assignment area. In reading the article, relate its contents to the faculty notes and readings, paying particular attention to the components of the experimental study design. As you read, think about the investigator's choices in: developing the study hypothesis selecting the treatment and comparison groups, and how they conducted randomization and blinding of treatment status Also, consider alternatives to their choices that may have improved the study or perhaps made it even more prone to bias. Questions to Consider: Think about the design of this randomized controlled trial. What was its purpose? What were the treatment and control assignments, and are they sufficient to answer the study hypothesis? Studies such as this are frequently used to directly inform clinical practice. How might a trial such as this one provide useful information for clinicians treating a patient with osteoarthritis? Perhaps most important, are both the efficacy of the drug and the risk of side effects sufficient to recommend it to patients? Step 2 After completing Step 1, answer the following questions to the best of your ability. You may want to record your answers in a Word document. Note: Step 2 will not be graded - it is for your practice and understanding: a. What were the study's rationale/hypothesis and objectives? b. Was there any way that the blinding of treatment status could have be compromised (in other words, the investigator somehow figures out treatment status)? If so, what might have been its potential impact? c. Do you see any problems with the way the outcome of interest was measured? If so, suggest some alternatives. d. Are there any possible concerns about patient compliance with their assigned treatment? What impact might noncompliance have on study findings? e. Why do you think there was an entire paragraph dedicated to disclosure of who conducted, oversaw, and analyzed the study's findings? What implications does this information have on the study? f. To what population(s) do you think this study's findings may or may not be generalizable, and why? g. This trial may have some relevance to clinical practice, but there were adverse events associated Page 1 Excelsior College PBH 321 with some of the treatment levels. Based on the frequency and type of adverse events described in the article, what advice might you want to provide physicians regarding the use of Tanezumab for osteoarthritis pain? Step 3 Read the commentary I have provided and compare your answers to my analysis. At this point, you may want to review the article to reinforce some of the points I have made in my critical examination of the research article. You will find a link to my commentary in the assignment area. Step 4 Now it is your turn to try analyzing epidemiological research! Proceed to read the article, \"Mussolino et al., Jogging and Bone Mineral Density in Men: Results From NHANES III, American Journal of Public Health, 2001; 10561059.\" You will find a link for this article in the assignment area. As you read the article, I would like you to relate the contents of the article to the faculty notes and readings, paying particular attention to the components of the crosssectional study design. Think about the strengths and limitations of the crosssectional study design. What are the problems that arise from using data collected for another purpose? Based on what you understand about the study design, what are the limitations of the investigator's conclusions? Consider how useful the crosssectional study design is when resources are limited and a tentative public health hypothesis needs to be evaluated. How does the availability of the NHANES database make conducting this type of study feasible? There are also some disadvantages to the crosssectional approach. Could this same study have been conducted using an experimental design? Use the example response for the research article by Lane et al. to guide you as you develop your answers. Step 5 After reading the article, answer the following questions by recording your answers in a Word document. As this is a different type of research study, the questions are slightly different from the ones you answered in the previous practice session. Note: your answers for this second study will be graded. Submit your Word document in the dropbox for this activity. Remember to answer these questions in your own words. Do not copy entire phrases from the article without attribution. a. What were the study's rationale/hypothesis and objectives? b. Was the use of an existing data source potentially problematic? What are the strengths and weaknesses? c. What were the inclusion/exclusion criteria for the study analyses? d. What was the outcome of interest? How was it measured, with respect to the timing of the exposure (jogging)? Can you think of any advantages/disadvantages of the way either the exposure or outcome was measured? e. To what population(s) do you think this study's findings may or may not be generalizable, and why? f. Are there any hypotheses that can be generated from the major findings? g. Could this same study have been conducted using an experimental design? Why or why not? Step 6 Once you have answered the questions in Step 5 and created a Word document, submit it to the dropbox for this activity. Page 2 Jogging and Bone Mineral Density in Men: Results From NHANES III A B S T R A C T Objectives. This cross-sectional population-based study assessed the association of jogging with femoral bone mineral density (BMD) in men. Methods. Data are from a nationally representative sample of 4254 men aged 20 to 59 years from the Third National Health and Nutrition Examination Survey (NHANES III). Total femoral BMD was measured by dual energy x-ray absorptiometry. Jogging was self-reported. Results. Jogging (any vs none) was strongly associated with higher BMD in multivariate models (P < .01) for both young and middle-aged men. Men who jogged 9 or more times per month had higher BMD levels than those who jogged only 1 to 8 times per month (P = .01). Conclusions. Jogging is associated with higher femoral neck BMD in men. Additional large-scale studies that measure all aspects of jogging are warranted. (Am J Public Health. 2001;91: 1056-1059) Michael E. Mussolino, MA, Anne C. Looker, PhD, and Eric S. Orwoll, MD Jogging is enjoyed by millions of people in the United States. Individuals who undertake high-impact activities like jogging on a regular basis may have higher bone mineral density (BMD), particularly at the femoral neck.1 Higher levels of weight-bearing or vigorous physical activity have been shown to be associated with a lower incidence of hip fractures.2,3 A number of studies have investigated the effect of jogging on BMD, with many focusing on women.4-9 Among men, the results for jogging and BMD have been inconsistent. Some studies have found higher BMD among joggers in the femoral neck, lumbar spine, or total body,1,10-12 while others have found an association only at lower limb sites.13,14 Other studies have found that BMD among runners is the same as or lower than in nonrunners.15,16 A problem of the studies of both men and women has been the use of self-selected athletes, which, coupled with small sample sizes, likely limits generalizability. Other analyses of BMD and physical activity have combined jogging with other strenuous activities.17 The Third National Health and Nutrition Examination Survey (NHANES III), a large-scale national study, provides an opportunity to investigate the relation between jogging and BMD in a representative sample of young and middleaged men with a wide variation in self-reported jogging frequency that likely encompasses the casual jogger as well as the more serious runner. Thus, our results may be more broadly generalizable than those of previous studies. Methods Data for these analyses were taken from NHANES III, which collected data from a national probability sample of the civilian noninstitutionalized US population from 1988 to 1994.18 Analyses were limited to 4603 nonHispanic White, non-Hispanic Black, and Mexican American men whose BMD levels were 1056 American Journal of Public Health measured and who were aged 20 to 59 years at the time of the NHANES III examination, because relatively few study respondents 60 years or older reported jogging. Men of other race/ ethnicity groups were omitted because of the small number of observations. Also excluded were 3 men with missing values on the jogging variables and 346 men with unknown values on any of the other variables assessed in the study. A total of 4254 men were included in the analysis (954 joggers and 3300 nonjoggers). Because nonjoggers in the main analysis may have participated in other leisure activities, we also did a subanalysis in which nonjoggers were restricted to those who reported no leisure activities (n=577). Measurement of BMD BMD was measured by trained examiners in mobile examination centers. Total femoral bone density was measured by dual energy x-ray absorptiometry (Hologic QDR1000; Hologic, Inc, Waltham, Mass). Scans were reviewed by consultants at the Mayo Clinic, Rochester, Minn, for quality control.19 Jogging Assessment For data on self-reported jogging, the following questions were asked: \"In the past month, did you jog or run?\" and \"In the past month, how often did you jog or run?\" Jogging was categorized as 9 or more times per month, 1 to 8 times per month, and none. Michael E. Mussolino and Anne C. Looker are with the National Center for Health Statistics, Centers for Disease Control and Prevention, Hyattsville, Md. Eric S. Orwoll is with Oregon Health Sciences University, Portland. Requests for reprints should be sent to Michael E. Mussolino, MA, Division of Epidemiology, National Center for Health Statistics, 6525 Belcrest Rd, Suite 730, Hyattsville, MD 20782 (e-mail: mmussolino@cdc.gov). This article was accepted January 12, 2001. July 2001, Vol. 91, No. 7 Other Baseline Variables Information on smoking status (current, former, or never), alcohol consumption (drinks per month), self-assessed health status, food energy, calcium consumption, protein intake, weight history (including maximum weight), and chronic conditions related to secondary osteoporosis were obtained by interview. We calculated body mass index (BMI) from measurements of the subjects' height and weight. The nutritional variables were determined on the basis of what the subject recalled eating in the past 24 hours. The chronic-conditions variable was based on self-reported doctor's diagnoses of congestive heart failure, stroke, chronic bronchitis, goiter, other thyroid disease, and diabetes. Statistical Analysis Statistical analyses were performed with linear regression procedures in SAS20 and SUDAAN.21 We used sample weights when calculating point estimates, so estimates are representative of the civilian noninstitutionalized US population at the time of NHANES III. Weighted multivariate regression analyses were performed for the total sample and stratified by age (20-39 and 40-59 years). All models were adjusted for age at interview, race/eth- nicity, BMI, food energy, calcium consumption, protein intake, smoking status, alcohol consumption, chronic conditions, and weight change, unless otherwise indicated. Results Baseline characteristics of men by jogging status (any vs none) are shown in Table 1. Joggers tended to be younger, so all subsequent comparisons were made after age was controlled for. Joggers weighed less than nonjoggers and were more likely to have never smoked, have no chronic conditions, and be in excellent health. Overall, 22.3% of men reported jogging in the past month. Mean femoral BMD was 5.0% higher among joggers than among nonjoggers; when joggers were compared with sedentary nonjoggers, the figure rose to 7.7% (1.069 g/cm2 vs 0.993 g/cm2). In the multivariate models, men who reported any jogging had a significantly higher BMD than those who did not jog (P<.001). Results of categorizing jogging status by frequency are shown in Table 2. Men who reported jogging 9 or more times per month or 1 to 8 times per month had higher BMD levels than men who did not jog (P<.01). Furthermore, men who jogged 9 or more times per month had significantly higher BMD lev- TABLE 1Characteristics of 4254 Male Joggers and Nonjoggers Aged 20 to 59 Yearsa: Third National Health and Nutrition Examination Survey Jogging Status Any N Age at interview, y Weight, kg Height, cm Body mass index, kg/m2 Total femoral bone mineral density, g/cm2 Smoking status, % Current Former Never Alcohol, drinks/mo Any chronic conditions, % Weight loss from maximum 10%, % Dietary variables Food energy, kcal/day Calcium, mg/day Protein, g/day Self-assessed health status, % Excellent Very good Good Fair Poor a Noneb 954 33.4* 81.4* 176.9 26.0* 1.069* 3300 38.2 83.2 176.8 26.6 1.018 23.2* 28.1 48.6* 11.6 4.2* 26.1 40.0 25.4 34.6 13.4 8.7 24.9 2840.7 1016.5 105.0 34.2* 39.3 21.7* 4.4* 0.4* 2828.5 1036.4 104.7 20.1 34.6 34.4 9.5 1.4 Controlling for age at interview. Of the 3300 nonjoggers, 2723 were \"active nonjoggers\" (i.e., reported other physical activity) and 577 were \"sedentary nonjoggers\" (i.e., did not report any physical activity). *P < .05 (t test). b July 2001, Vol. 91, No. 7 els than those who jogged 1 to 8 times per month (P=.01) (data not shown). Age-specific analyses revealed similar results for ages 20 to 39, while only the category of 9 or more times per month was significant for ages 40 to 59. The relatively small number of joggers in this age group may have led to the reduced P values in the category of 1 to 8 times per month. The interaction of race/ethnicity and jogging group was not significant (P=.56), suggesting that the effect of jogging frequency is the same in all 3 race/ethnicity groups. In a subanalysis to assess the effect of jogging frequency on BMD, we found that the BMD levels of those who jogged frequently (i.e., >20 times/month) were similar to the levels of those who jogged less frequently, suggesting that there may be a ceiling for the graded effect beyond which no further benefit occurs. When treated as a continuous variable in the multivariate analyses, jogging remained a strong predictor of higher BMD (P<.01) (data not shown). When the self-assessed health status variable was added to the regression models, jogging remained a statistically significant predictor of BMD (P<.01). We also investigated the impact of total number of leisure activities on the relationship between jogging and BMD. Leisure activities varied, but walking and gardening/yard work were the most reported activities among both joggers (68.0% and 61.4%, respectively) and physically active nonjoggers (56.8% and 66.1%, respectively). Joggers reported a larger number of leisure activities than nonjoggers (mean = 4.8 vs 2.6). Joggers also reported a greater number of weight-bearing activities (mean=2.9 vs 1.1). In regression models adjusted for age and total number of leisure activities, BMD remained significantly higher for joggers (P<.01). Results were similar after age and number of weight-bearing activities were adjusted for (P<.01). Discussion Our results from this large nationally representative sample are consistent with previous research suggesting that jogging is associated with higher femoral BMD in men. Adjustment for established BMD risk factors did not appreciably alter the conclusions. Furthermore, among the 954 joggers in the study, 47.3% reported running 8 or fewer times per month, suggesting that even infrequent jogging may be beneficial to BMD. In our study, we were able to examine one aspect of jogging on BMD: frequency. Other aspects of jogging, such as distance or pace, could not be assessed because NHANES III did not include questions on these items. To American Journal of Public Health 1057 TABLE 2Frequency of Jogging and Bone Mineral Density (BMD) for Men Aged 20 to 59 Yearsa: Third National Health and Nutrition Examination Survey Jogging Frequency Ages 20-39 9 times/mo 1-8 times/mo None Ages 40-59 9 times/mo 1-8 times/mo None Ages 20-59 9 times/mo 1-8 times/mo None n BMD (g/cm2) P 350 405 1796 1.113 1.083 1.043 .0705 .0401 (Reference) .0000 .0015 (Reference) 98 101 1504 1.076 1.034 1.011 .0651 .0236 (Reference) .0021 .1487 (Reference) 448 506 3300 1.104 1.071 1.036 .0685 .0354 (Reference) .0000 .0010 (Reference) a Models were adjusted for age at interview, race/ethnicity, body mass index, food energy, calcium consumption, protein intake, smoking status, alcohol consumption, chronic conditions, and weight change. our knowledge, jogging pace has not been examined in previous studies. Several studies have examined the effect of jogging distance on BMD in men, and all but one report22 suggested that there may be a ceiling beyond which additional distance does not improve BMD. For example, one study reported that men who ran more than 64 km per week had significantly lower vertebral BMD than nonrunners,23 and another concluded that male long-distance runners had reduced BMD and increased bone turnover compared with controls.16 MacDougall et al.14 found no further increase in lower-leg BMD in male runners who ran more than 20 miles per week compared with those who ran less; in fact, increased distance was potentially detrimental (i.e., mean BMD in those who ran 60-75 miles per week was similar to that in controls). Our results showing similar femoral BMD in those who jogged frequently (i.e., >20 times/month) and those who jogged less frequently, while exploratory in nature, were consistent with these studies. The mechanism underlying this ceiling effect is probably multifactorial and may include factors such as body weight14 or sex hormone status.24 Our study focused on the proximal femur, which is a skeletal site that receives increased loading from jogging. The effect of jogging on other skeletal sites, which may not be as directly loaded, has varied. One longitudinal study found significantly higher lumbar spine BMD levels in older male runners,11 but other studies have found similar lumbar spine BMD levels in runners and controls.1,15 Running also does not appear to affect BMD in the forearm.3 Our study has the advantage of including joggers with widely diverse jogging frequency. Thus, our results may be more applicable to the general population than those of previous studies, which tended to focus on more dedi1058 American Journal of Public Health cated or elite athletes. However, our study has several limitations. It is cross-sectional in nature and therefore cannot provide definitive evidence that jogging caused the higher BMD levels observed in male joggers. The dose- response seen for jogging frequency is suggestive in this regard, but it is also possible that self-selection played a role (i.e., those who chose to jog had higher BMD before commencing to jog). Results did not change when we included in the analysis self-reported health status or the presence of chronic conditions, which suggests that the relationship between jogging and BMD is not due to differences in health status between joggers and nonjoggers. However, self-selection could also occur because those who choose to jog are better adapted physically to perform this activity (i.e., their musculoskeletal system, including BMD, is better suited to jogging). This would not be addressed by the overall health status or chronic condition variables. We also used a self-reported assessment of jogging in our study rather than a direct measurement. However, other data suggest that self-reported physical activity is reasonably valid. For example, in the predictive models of cardiorespiratory fitness that also included weight, resting heart rate, and current smoking, self-reported physical activity was found to be the principal contributor.25 The associations consistently found between self-reported physical activity and lower risk of several diseases also support the validity of the measure.25 Our study looked at leisure-time physical activity only, so it could underestimate total physical activity. However, most people do not have jobs requiring regular physical exertion.25 We conclude that jogging is associated with significantly higher femoral BMD in men. This effect was graded in nature and was ob- served in both younger and middle-aged men. This finding may have public health significance, since femoral BMD is a strong predictor of hip fracture, which is the most devastating consequence of osteoporosis from a public health standpoint. Most important, our data suggest that even a modest frequency of jogging is associated with higher femoral BMD and thus may provide some protection against osteoporosis and fracture. Additional largescale studies that measure all aspects of jogging are warranted. Contributors M. E. Mussolino was primarily responsible for the study design, writing, and data analyses. A.C. Looker provided input into the initial study design, wrote parts of the \"Discussion\" section, and revised other parts of the manuscript. E. S. Orwoll provided guidance on interpreting data and contributed to the analysis of the impact of other leisure activities on the relationship between jogging and BMD. All authors participated in preparing the manuscript. References 1. Brahm H, Strom H, Piehl-Aulin K, Mallmin H, Ljunghall S. Bone metabolism in endurance trained athletes: a comparison to populationbased controls based on DXA, SXA, quantitative ultrasound, and biochemical markers. Calcif Tissue Int. 1997;61:448-454. 2. Kujala UM, Kaprio J, Kannus P, Sarna S, Koskenvuo M. Physical activity and osteoporotic hip fracture risk in men. Arch Intern Med. 2000; 160:705-708. 3. Joakimsen RM, Magnus JH, Fonnebo V. Physical activity and predisposition for hip fractures: a review. Osteoporos Int. 1997;7:503-513. 4. Petit MA, Prior JC, Barr SI. Running and ovulation positively change cancellous bone in premenopausal women. Med Sci Sports Exerc. 1999;31:780-787. 5. Emslander HC, Sinaki M, Muhs JM, et al. Bone mass and muscle strength in female college athletes (runners and swimmers). Mayo Clin Proc. 1998;73:1151-1160. 6. Taaffe DR, Robinson TL, Snow CM, Marcus R. High-impact exercise promotes bone gain in well-trained female athletes. J Bone Miner Res. 1997;12:255-260. 7. Etherington J, Harris PA, Nandra D, et al. The effect of weight-bearing exercise on bone mineral density: a study of female ex-elite athletes and the general population. J Bone Miner Res. 1996;11:1333-1338. 8. Snow-Harter C, Bouxsein ML, Lewis BT, Carter DR, Marcus R. Effects of resistance and endurance exercise on bone mineral status of young women: a randomized exercise intervention trial. J Bone Miner Res. 1992;7:761-769. 9. Jonsson B, Ringsberg K, Josefsson PO, Johnell O, Birch-Jensen M. Effects of physical activity on bone mineral content and muscle strength in women: a cross-sectional study. Bone. 1992;13: 191-195. 10. Need AG, Wishart JM, Scopacasa F, Horowitz M, Morris HA, Nordin BEC. Effect of physical July 2001, Vol. 91, No. 7 11. 12. 13. 14. 15. activity on femoral bone density in men. BMJ. 1995;310:1501-1502. Lane NE, Oehlert JW, Bloch DA, Fries JF. The relationship of running to osteoarthritis of the knee and hip and bone mineral density of the lumbar spine: a 9 year longitudinal study. J Rheumatol. 1998;25:334-341. Stewart AD, Hannan J. Total and regional bone density in male runners, cyclists, and controls. Med Sci Sports Exerc. 2000;32:1373-1377. Bennell KL, Malcolm SA, Khan KM, et al. Bone mass and bone turnover in power athletes, endurance athletes and controls: a 12-month longitudinal study. Bone. 1997;20:477-484. MacDougall JD, Webber CE, Martin J, et al. Relationship among running mileage, bone density, and serum testosterone in male runners. J Appl Physiol. 1992;73:1165-1170. Goodpaster BH, Costill DL, Trappe SW, Hughes GM. The relationship of sustained exercise train- July 2001, Vol. 91, No. 7 16. 17. 18. 19. 20. ing and bone mineral density in aging male runners. Scand J Med Sci Sports. 1996;6:216-221. Hetland ML, Haarbo J, Christiansen C. Low bone mass and high bone turnover in male long distance runners. J Clin Endocrinol Metab. 1993;77:770-775. Greendale GA, Barrett-Connor E, Edelstein S, Ingles S, Haile R. Lifetime leisure exercise and osteoporosis. The Rancho Bernardo Study. Am J Epidemiol. 1995;141:951-959. Plan and operation of the Third National Health and Nutrition Examination Survey, 1988-94. National Center for Health Statistics. Vital Health Stat 1. 1994;No. 32. Wahner HW, Looker A, Dunn WL, Hauser MF, Walters LC, Novak C. Quality control of bone densitometry in a national health survey (NHANES III) using three mobile examination centers. J Bone Miner Res. 1994;9:951-960. SAS/STAT User's Guide, Version 6. 4th ed. Cary, NC: SAS Institute Inc; 1990. 21. Shah BV, Barnwell BG, Bieler GS. SUDAAN User's Manual: Software for Analysis of Correlated Data, Release 6.40. Research Triangle Park, NC: Research Triangle Institute; 1995. 22. Lane NE, Bloch DA, Jones HH, Marshall WH Jr, Wood PD, Fries JF. Long-distance running, bone density, and osteoarthritis. JAMA. 1986; 255:1147-1151. 23. Bilanin JE, Blanchard MS, Russek-Cohen E. Lower vertebral bone density in male long distance runners. Med Sci Sports Exerc. 1989;21: 66-70. 24. Cooper CS, Taaffe DR, Guido D, Packer E, Holloway L, Marcus R. Relationship of chronic endurance exercise to the somatotropic and sex hormone status of older men. Eur J Endocrinol. 1998;138:517-523. 25. Physical Activity and Health: A Report of the Surgeon General. Atlanta, Ga: National Center for Chronic Disease Prevention and Health Promotion; 1996. American Journal of Public Health 1059 The n e w e ng l a n d j o u r na l of m e dic i n e original article Tanezumab for the Treatment of Pain from Osteoarthritis of the Knee Nancy E. Lane, M.D., Thomas J. Schnitzer, M.D., Ph.D., Charles A. Birbara, M.D., Masoud Mokhtarani, M.D., David L. Shelton, Ph.D., Mike D. Smith, Ph.D., and Mark T. Brown, M.D. A BS T R AC T Background Increased expression of nerve growth factor in injured or inflamed tissue is associated with increased pain. This proof-of-concept study was designed to investigate the safety and analgesic efficacy of tanezumab, a humanized monoclonal antibody that binds and inhibits nerve growth factor. Methods We randomly assigned 450 patients with osteoarthritis of the knee to receive tanezumab (administered at a dose of 10, 25, 50, 100, or 200 g per kilogram of body weight) or placebo on days 1 and 56. The primary efficacy measures were knee pain while walking and the patient's global assessment of response to therapy. We also assessed pain, stiffness, and physical function using the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC); the rate of response using the criteria of the Outcome Measures for Rheumatology Committee and Osteoarthritis Research Society International Standing Committee for Clinical Trials Response Criteria Initiative (OMERACT-OARSI); and safety. Results When averaged over weeks 1 through 16, the mean reductions from baseline in knee pain while walking ranged from 45 to 62% with various doses of tanezumab, as compared with 22% with placebo (P<0.001). Tanezumab, as compared with placebo, was also associated with significantly greater improvements in the response to therapy as assessed with the use of the patients' global assessment measure (mean increases in score of 29 to 47% with various doses of tanezumab, as compared with 19% with placebo; P0.001). The rate of response according to the OMERACT-OARSI criteria ranged from 74 to 93% with tanezumab treatment, as compared with 44% with placebo (P<0.001). The rates of adverse events were 68% and 55% in the tanezumab and placebo groups, respectively. The most common adverse events among tanezumab-treated patients were headache (9% of the patients), upper respiratory tract infection (7%), and paresthesia (7%). From the University of California at Davis Medical School, Sacramento (N.E.L.); Northwestern University Feinberg School of Medicine, Chicago (T.J.S.); University of Massachusetts School of Medicine, Worcester (C.A.B.); Rinat Neuroscience, South San Francisco, CA (M.M., D.L.S.); and Pfizer, New London, CT (M.D.S., M.T.B.). Address reprint requests to Dr. Lane at the Department of Medicine, University of California at Davis Medical School, 4800 Second Ave., Suite 2600, Sacramento, CA 95817, or at nelane@ucdavis.edu. Drs. Lane and Schnitzer contributed equally to this article. This article (10.1056/NEJMoa0901510) was published on September 29, 2010, at NEJM .org. N Engl J Med 2010;363:1521-31. Copyright 2010 Massachusetts Medical Society. Conclusions In this proof-of-concept study, treatment with tanezumab was associated with a reduction in joint pain and improvement in function, with mild and moderate adverse events, among patients with moderate-to-severe osteoarthritis of the knee. (Funded by Rinat Neuroscience; ClinicalTrials.gov number, NCT00394563.) n engl j med 363;16 nejm.org october 14, 2010 1521 The New England Journal of Medicine Downloaded from nejm.org at NEW YORK STATE LIB ACQUISITIONS on January 9, 2011. For personal use only. No other uses without permission. Copyright 2010 Massachusetts Medical Society. All rights reserved. The N n e w e ng l a n d j o u r na l erve growth factor is a neurotro phin that regulates the structure and function of responsive sensory neurons, including small-diameter nociceptive afferents. There has been increasing recognition of the potential role of nerve growth factor in pain modulation through nociceptor sensitization.1-7 In animals and humans, exogenous nerve growth factor increases pain either locally or systemically, depending on the dose and the route of administration.8,9 Increased expression of nerve growth factor is found in inflamed tissues from patients with conditions such as arthritis, pancreatitis, and prostatitis.10-12 Levels of nerve growth factor are also elevated in animal models of inflammatory pain, and pharmacologic inhibition of the activity of nerve growth factor in these models reduces or blocks signs of pain. Therefore, nerve growth factor appears to have a role in causing and augmenting pain in these models.1,13-15 The development of therapeutic interventions that are based on antagonism of nerve growth factor is of interest.16,17 The treatment options for patients with painful osteoarthritis of the knee are inadequate. Nonsteroidal antiinflammatory drugs and narcotic analgesics are commonly used18,19; however, these medications have well-described gastrointestinal and cardiorenal side effects,20,21 and the response to them is unsatisfactory in some patients.22,23 Potent analgesic medications with acceptable sideeffect profiles may help to avoid or delay surgical intervention.24 Tanezumab is a humanized IgG2 monoclonal antibody directed against nerve growth factor that blocks the interaction of nerve growth factor with its receptors, TrkA and p75.25 A small phase 1 clinical trial showed that a single intravenous injection of tanezumab substantially reduced pain in patients with osteoarthritis of the knee.26,27 We report the results of a proof-of-concept study of tanezumab in patients with advanced osteoarthritis of the knee who did not have a satisfactory response to nonopiate pain medications or who were considered to be candidates for invasive intervention. We compared the safety, side-effect profile, and efficacy of repeat doses of tanezu mab as compared with placebo. of m e dic i n e the basis of American College of Rheumatology criteria,28 with radiographic confirmation (Kellgren-Lawrence grade 2 or higher, on a scale of 0 to 4, with higher numbers indicating more severe signs of osteoarthritis). Patients were eligible only if they were unwilling to take nonopiate pain medications or had had an unsatisfactory response to them or if they were candidates for or seeking invasive interventions such as intraarticular injections or total knee replacement. All pain medications except the \"rescue\" medications, acetaminophen and tramadol, were discontinued at the screening visit. At the time of randomization, patients had to have pain while walking on a flat surface (the walking-pain measure of the Western Ontario and McMaster Universities Osteoarthritis Index [WOMAC]) that they rated between 50 and 90 on a visual-analogue scale that ranged from 0 to 100, with 100 indicating maximal pain. In addition, among patients who discontinued pain medication during the screening period, an increase in the walking-pain score of 10 or more was required between screening and randomization. The exclusion criteria were pregnancy, a history of or current symptoms of an autoimmune disorder, cancer within the previous 5 years except for cutaneous basal-cell or squamous-cell cancer resolved by excision, allergic reaction to monoclonal antibodies or IgG-fusion proteins, infection with hepatitis B or hepatitis C virus or the human immunodeficiency virus, drug abuse, fibromyalgia, clinically significant cardiac disease, diabetes mellitus requiring oral treatment or insulin, clinically significant neurologic disease, or a clinically significant psychiatric disorder. All participants provided written informed consent. Study Design and Oversight Patients were recruited between March 30, 2006, and May 3, 2007, at 46 study centers in the United States and were screened within 30 days before randomization. Eligible patients who were taking pain medication other than acetaminophen and tramadol underwent a washout period (of at least 5 half-lives of the medication). Patients rated their knee pain and recorded the score in an electronic diary every day for 3 days before randomization to establish their baseline pain score. Eligible patients were randomly assigned on day 1, with the Me thods use of an interactive voice-response system, to plaStudy Population cebo or to tanezumab at a dose of 10, 25, 50, 100, We enrolled patients, 40 to 75 years of age, who or 200 g per kilogram of body weight, such that had osteoarthritis of the knee as diagnosed on there were equal numbers in each study group. 1522 n engl j med 363;16 nejm.org october 14, 2010 The New England Journal of Medicine Downloaded from nejm.org at NEW YORK STATE LIB ACQUISITIONS on January 9, 2011. For personal use only. No other uses without permission. Copyright 2010 Massachusetts Medical Society. All rights reserved. Tanezumab for Osteoarthritis of the Knee A pharmacist at each study site received each patient's randomization number and prepared each patient's dosing solution. Other than the pharmacist at each site and one statistician at the contract research organization, all the staff members and patients involved in the study were unaware of the group assignments. The study medication was administered intravenously on days 1 and 56. Study visits were scheduled for days 14, 28, 70, 84, 112, 136, and 182, during which safety and efficacy assessments were performed and serum samples for routine laboratory tests and for pharmacokinetic analyses were obtained. In addition, patients were contacted by telephone on day 42 to ask about adverse events. Patients recorded their knee pain and their use of rescue medication daily in an electronic diary. The rescue medications that were permitted were acetaminophen at a dose of 3000 mg or less per day, tramadol at a dose of 400 mg or less per day, or both, during the washout period and days 1 through 28 (weeks 1 through 4) and acetaminophen at a dose of 3000 mg per day for the remainder of the study. Patients could enter an open-label extension of the trial (ClinicalTrials.gov number, NCT00399490) at week 16 (day 112) if they had received two doses of the study drug and had been followed for at least 8 weeks after the last dose. The study was designed and coordinated by Rinat Neuroscience, a subsidiary of Pfizer. An external data and safety monitoring board monitored safety, and day-to-day study operations, including data management, were overseen by PPDI (a contract research organization contracted by Pfizer). The data were analyzed by Pfizer. The first author wrote the first draft of the manuscript. All the authors were involved in the design of the study and interpretation of the data, contributed to the writing of the manuscript, made the decision to submit the manuscript for publication, attest that the study was performed in accordance with the protocol and the statistical analysis plan, and vouch for the accuracy and completeness of the reported results. Editorial support was provided by UBC Scientific Solutions and was funded by Pfizer. The study protocol was approved by the local ethics committee at each study center before patient enrollment began. The protocol, including the statistical analysis plan, is available with the full text of this article at NEJM.org. Efficacy Assessments The primary efficacy outcomes were the change from baseline in the pain the patient felt in the index knee while walking on a flat surface and in the patient's global assessment of response to therapy, averaged over weeks 1 through 16. Secondary efficacy outcomes included the change from baseline in overall knee pain and in scores on the WOMAC subscales for pain, stiffness, and physical function. Pain while walking and overall knee pain were recorded daily in an electronic diary, whereas the patient's global assessment of response to therapy and scores on the WOMAC subscales were recorded on study-visit days. Pain, the patient's global assessment, and scores on the WOMAC subscales were assessed with the use of a visual-analogue scale that ranged from 0 to 100. In the case of pain and WOMAC scores, a lower score indicated improvement (i.e., less pain, less stiffness, and less limitation of physical function), whereas in the case of the patient's global assessment, a higher score indicated improvement (i.e., a better response to therapy). Another secondary outcome was the response to therapy on the basis of the criteria of the Outcome Measures for Rheumatology Committee and Osteoarthritis Research Society International Standing Committee for Clinical Trials Response Criteria Initiative (OMERACT-OARSI).29 Patients were classified as having had a response if the WOMAC pain or physical-function score decreased by 50% or more and by 20 or more points on the visual-analogue scale or if two of the following three findings were recorded: a decrease in the WOMAC pain score by 20% or more and by 10 or more points on the visual-analogue scale, a decrease in the WOMAC physical-function score by 20% or more and by 10 or more points on the scale, or an increase in the score on the patient's global assessment by 20% or more and by 10 or more points on the scale. Rescue medication use, also a prespecified secondary outcome, was recorded daily in the patients' diaries. Safety Assessments The nature, onset, duration, severity, and outcome of all adverse events, as well as any relationship of an adverse event to the study drug were ascertained and documented at each visit. Safety assessments included physical and neurologic examinations (e.g., evaluation of mental status, n engl j med 363;16 nejm.org october 14, 2010 1523 The New England Journal of Medicine Downloaded from nejm.org at NEW YORK STATE LIB ACQUISITIONS on January 9, 2011. For personal use only. No other uses without permission. Copyright 2010 Massachusetts Medical Society. All rights reserved. The n e w e ng l a n d j o u r na l of m e dic i n e medication and underwent at least one efficacy assessment (the modified intention-to-treat population) (Fig. 1). Most patients in the modified intention-to-treat population had a Kellgren-Lawrence grade of 3 (52%) or 4 (17%) and severe pain33 both at the time of screening (mean [SD] score on the visual-analogue scale, 5812) and at the time of randomization, after washout of previous medications (score on the visual-analogue scale, 7111). A total of 87% of the patients reported Statistical Analysis taking pain medication for knee pain before they We estimated that we would need to enroll 75 enrolled in the study. The baseline characteristics patients in each group for the study to have 80% of the patients were similar across study groups power to detect a difference between the tanezu (Table 1). mab groups and the placebo group of 15 points or more on the visual-analogue scale for the pri- Efficacy mary outcomes (the average change from base- As compared with placebo, tanezumab, at all the line through week 16 in knee pain while walking doses studied, was associated with an improveand in the patient's global assessment of re- ment in the primary efficacy measures. The sponse to therapy), with an effect size of 0.5 (in- mean reduction from baseline in the score on the dicating a moderate difference).31 With respect to visual-analogue scale for knee pain while walkthe intensity of pain, decreases of 10 or more ing, averaged over weeks 1 through 16, ranged points on a visual-analogue scale that ranges from 31.0 to 45.2 points with various doses of from 1 to 100 are considered to be minimally tanezu ab, as compared with 15.5 points with m important improvements, and decreases of 20 or placebo (a reduction of 45 to 62% with tanezummore points are considered to be moderately im- ab vs. 22% with placebo, P<0.001 for the comportant improvements.32 parison of all doses of tanezumab with placebo) Changes from baseline in all the measures that (Fig. 2A). We observed significant improvements were assessed with the use of a visual-analogue among patients receiving tanezumab as comscale were determined with a mixed-model, re- pared with those receiving placebo by the end of peated-measures analysis, with model terms for the first week, and significant improvements study site, study group, study week, and the in- continued to be seen throughout the remainder teraction between study medication and study of the treatment period. The mean increase from week, and with the baseline score on the visual- baseline in the score on the patient's global assessanalogue scale as a covariate, with no imputation ment of response to therapy, averaged over weeks for missing data. A repeated-measures analysis 1 through 16, ranged from 16.3 to 23.7 points was also used to assess the number of rescue with various doses of tanezumab, as compared medication pills taken. For the analysis of rates with 9.2 points with placebo (an increase of 29 to of response according to OMERACT-OARSI crite- 47% with tanezumab vs. 19% with placebo, ria (calculated on the basis of the average change P0.001 for the comparison of all doses of tanfrom baseline to week 16), we used the Cochran- ezumab with placebo) (Fig. 2B). By week 2, the Mantel-Haenszel test, stratified according to study scores on the patient's global assessment had imsite, to compare the proportions of patients in the proved in the group receiving 25 g of tanezumtanezumab groups who had a response with the ab per kilogram, as compared with placebo proportion of those in the placebo group who had (P=0.002); by week 4, the scores had improved in a response. the groups receiving 50 g, 100 g, and 200 g of tanezumab per kilogram, as compared with placebo (P=0.01, P<0.001, and P<0.001 for the three R e sult s comparisons, respectively); and during weeks 10 Baseline Characteristics of the Patients and 12, the scores improved in the group receivOf the 450 patients who underwent randomiza- ing tanezumab at a dose of 10 g per kilogram, tion, 440 received at least one dose of the study as compared with placebo (P=0.008). The im- strength, reflexes, sensation, and coordination), cognitive testing with the use of the Hopkins Verbal Learning Test-Revised,30 assessment of postural vital signs, and electrocardiography. Clinically significant abnormalities on a neurologic examination performed by the investigator or adverse events suggestive of peripheral neuropathy were further evaluated by an independent neurologist. 1524 n engl j med 363;16 nejm.org october 14, 2010 The New England Journal of Medicine Downloaded from nejm.org at NEW YORK STATE LIB ACQUISITIONS on January 9, 2011. For personal use only. No other uses without permission. Copyright 2010 Massachusetts Medical Society. All rights reserved. Tanezumab for Osteoarthritis of the Knee 450 Patients underwent randomization 75 Were assigned to receive placebo 75 Were assigned to receive tanezumab, 10 g/kg 75 Were assigned to receive tanezumab, 25 g/kg 75 Were assigned to receive tanezumab, 50 g/kg 75 Were assigned to receive tanezumab, 100 g/kg 75 Were assigned to receive tanezumab, 200 g/kg 65 Discontinued study by wk 26 24 Entered OLE 28 Had no response 13 Had other reason 10 Completed study 56 Discontinued study by wk 26 36 Entered OLE 6 Had an adverse event 7 Had no response 7 Had other reason 19 Completed study 53 Discontinued study by wk 26 34 Entered OLE 1 Had an adverse event 12 Had no response 6 Had other reason 22 Completed study 62 Discontinued study by wk 26 38 Entered OLE 4 Had an adverse event 10 Had no response 10 Had other reason 13 Completed study 50 Discontinued study by wk 26 33 Entered OLE 3 Had an adverse event 4 Had no response 10 Had other reason 25 Completed study 60 Discontinued study by wk 26 38 Entered OLE 8 Had an adverse event 5 Had no response 9 Had other reason 15 Completed study 73 Were included in modified intentionto-treat analysis 74 Were included in safety analysis 74 Were included in modified intentionto-treat analysis 74 Were included in safety analysis 75 Were included in modified intentionto-treat analysis 74 Were included in safety analysis 72 Were included in modified intentionto-treat analysis 74 Were included in safety analysis 74 Were included in modified intentionto-treat analysis 74 Were included in safety analysis 72 Were included in modified intentionto-treat analysis 74 Were included in safety analysis Figure 1. Randomization and Follow-up. Eligible patients could enter the open-label extension (OLE) of the study at week 16. provements were maintained through week 16 in the groups receiving 25 g, 100 g, and 200 g of tanezumab per kilogram. The mean reductions from baseline in overall knee pain over the course of weeks 1 through 16 were similar in magnitude to those reported for knee pain while walking (reductions of 43 to 62% with tanezumab vs. 23% with placebo, P<0.001 for the comparison of all doses of tanezumab with placebo). Treatment with tanezumab, as compared with placebo, was also associated with reductions in the mean WOMAC scores for pain (reductions of 46 to 64% vs. 23%), stiffness (48 to 65% vs. 22%), and physical function (47 to 65% vs. 22%) over the same period (P<0.001 for all comparisons) (Table 2). The percentage of patients who had a response to therapy according to OMERACT-OARSI criteria, averaged over weeks 1 through 16, was significantly higher with tanezumab treatment than with placebo (74 to 93% vs. 44%, P<0.001 for the comparison of all doses of tanezumab with placebo) (Table 2). Rescue medications that were allowed per protocol were used less frequently by tanezumab-treated patients than by placebo-treated patients during weeks 1 through 16 (odds ratio with tanezumab, 0.50; 95% confidence interval [CI], 0.24 to 1.02; P=0.05) and was significantly lower during weeks 1 through 4 (odds ratio, 0.49; 95% CI, 0.24 to 0.99; P=0.04). Safety Among patients in the tanezumab groups, the three most common adverse events were headache, upper respiratory tract infection, and paresthesia (Table 3). The incidence of treatmentrelated adverse events was higher among patients treated with 100 g or 200 g of tanezumab per kilogram than among patients who received lower doses (28% and 35% in the groups receiving 100 g and 200 g per kilogram, respectively, vs. 11 to 18% in the groups receiving other doses). Peripheral sensory symptoms, including paresthesia, were reported in 14% of the patients receiving tanezumab and in 4% of those receiving placebo (Table 3); the severity of these adverse n engl j med 363;16 nejm.org october 14, 2010 1525 The New England Journal of Medicine Downloaded from nejm.org at NEW YORK STATE LIB ACQUISITIONS on January 9, 2011. For personal use only. No other uses without permission. Copyright 2010 Massachusetts Medical Society. All rights reserved. The n e w e ng l a n d j o u r na l of m e dic i n e Table 1. Baseline Characteristics of the Study Patients.* Placebo (N=74) Age yr Tanezumab, 25 g/kg (N=74) Tanezumab, 50 g/kg (N=74) Tanezumab, 100 g/kg (N=74) Tanezumab, 200 g/kg (N=74) 58.17.7 Characteristic Tanezumab, 10 g/kg (N=74) 58.38.3 59.98.1 60.47.7 57.18.2 58.47.6 Female sex no. (%) 42 (57) 49 (66) 50 (68) 37 (50) 44 (59) 40 (54) White race no. (%) 66 (89) 62 (84) 67 (91) 66 (89) 67 (91) 64 (86) 2 18/73 (25) 21/73 (29) 23/74 (31) 29/74 (39) 22/74 (30) 19/73 (26) 3 or 4 Kellgren-Lawrence grade no./total no. (%) 55/73 (75) 52/73 (71) 51/74 (69) 45/74 (61) 52/74 (70) 54/73 (74) Knee pain while walking 71.610.0 70.610.9 71.710.5 68.110.2 71.111.0 72.411.5 Patient's global assessment of response 48.820.8 55.720.3 51.020.6 51.616.9 49.919.9 54.422.4 WOMAC score Pain 69.011.9 65.813.9 69.212.5 62.112.3 68.313.2 68.412.0 Stiffness 74.413.5 69.713.1 75.012.4 66.717.5 71.217.9 73.313.1 Physical function 69.012.5 63.813.6 69.214.6 62.612.3 67.414.8 67.814.0 *\tPlus-minus values are means SD. \tRace was self-reported. \tA Kellgren-Lawrence score of 2 (minimal signs of osteoarthritis) indicates definite osteophytes without reduction of the joint space; a score of 3 (moderate signs of osteoarthritis) indicates diminished joint space; and a score of 4 (severe signs of osteoarthritis) indicates greatly reduced joint space. Data are from the intention-to-treat population, and missing data are excluded. Knee pain while walking was assessed with the use of a visual-analogue scale that ranged from 0 to 100, with higher scores indicating more pain. Data are from the modified intention-to-treat population. \tPatients' global assessment of response to therapy was assessed with the use of a visual-analogue scale that ranged from 0 to 100, with higher scores indicating a better response to therapy. Data are from the modified intention-to-treat population. \tScores on the pain, stiffness, and physical-function subscales of the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) were assessed with the use of a visual-analogue scale that ranged from 0 to 100, with higher scores indicating more pain, more stiffness, and more limitation of physical function, respectively. Data are from the modified intention-to-treat population. events was mild in 56 of the 70 patients with these symptoms in the tanezumab groups and in all 3 patients with these symptoms in the placebo group and was moderate in the remaining 14 patients in the tanezumab groups. The results of neurologic examinations of these patients were predominantly normal; in patients with clinically significant changes, abnormalities were minor and consisted mainly of changes in sensation in the extremities and diminished deep-tendon reflexes. A total of 15 patients had abnormal peripheral sensation that was unresolved at the final visit. Of these 15 patients, 6 had clinically significant findings on neurologic examination: decreased ankle reflexes in 1 patient who was receiving 100 g of tanezumab per kilogram; decreased temperature and a sensation of sharp pain, vibration, or both, in the toes or feet in 4 patients (1 each in the groups receiving 25 and 100 g of tanezumab per kilogram and 2 in the 1526 group receiving 200 g per kilogram); and bilateral decreased sensation in pain, fine touch, and temperature in a radial-nerve distribution in 1 patient who was receiving 100 g of tanezumab per kilogram; all other aspects of the neurologic examination were normal. Additional follow-up data were available for 7 patients with abnormal peripheral sensation who subsequently participated in the open-label extension study, and in each case, the adverse event resolved before the patient's entry into the extension study. Similarly, follow-up data were available for 1 of the 6 patients with neurologic findings, and in the case of this patient, the neurologic findings had normalized by the time of the baseline visit in the open-label extension study. The mean time to the onset of abnormal sensory symptoms was 33 days (median, 14) after the first dose of tanezumab, and the mean duration of symptoms was 18 days (median, 4) across tan- n engl j med 363;16 nejm.org october 14, 2010 The New England Journal of Medicine Downloaded from nejm.org at NEW YORK STATE LIB ACQUISITIONS on January 9, 2011. For personal use only. No other uses without permission. Copyright 2010 Massachusetts Medical Society. All rights reserved. Tanezumab for Osteoarthritis of the Knee A Patient's Assessment of Knee Pain while Walking Dose 1 Mean Change from Baseline over Weeks 1-16 Dose 2 0 Placebo 15.52.6 Tanezumab, 10 g/kg 32.12.5 Tanezumab, 25 g/kg 10 36.02.5 42.52.5 Tanezumab, 200 g/kg 20 31.02.6 Tanezumab, 100 g/kg Change from Baseline Tanezumab, 50 g/kg 45.22.6 30 40 50 60 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Week B Patient's Global Assessment of Response Dose 1 Mean Change from Baseline over Weeks 1-16 Dose 2 Placebo 9.21.8 Tanezumab, 10 g/kg Tanezumab, 25 g/kg 30 16.31.7 23.61.6 17.51.7 Tanezumab, 100 g/kg Change from Baseline Tanezumab, 50 g/kg 23.71.6 Tanezumab, 200 g/kg 21.01.7 20 10 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Week Figure 2. Change from Baseline in Patients' Assessment of Knee Pain while Walking and in Patients' Global Assessment of Response to Therapy. The patient's assessment of knee pain while walking and the patient's global assessment of response to therapy were obtained at baseline and at the indicated times with the use of a visual-analogue scale that ranged from 0 to 100. In the case of knee pain, a decrease in the score indicates improvement (i.e., less pain); in the case of the patient's global assessment, an increase in the score indicates improvement (i.e., a better response to therapy). Changes are reported as least-squares means SE. P<0.001 for the comparisons of all doses of tanezumab with placebo in the assessment of knee pain and global assessment of response, except for the comparison of 10 g of tanezumab per kilogram of body weight with placebo in the patient's global assessment, for which P=0.001. n engl j med 363;16 nejm.org october 14, 2010 1527 The New England Journal of Medicine Downloaded from nejm.org at NEW YORK STATE LIB ACQUISITIONS on January 9, 2011. For personal use only. No other uses without permission. Copyright 2010 Massachusetts Medical Society. All rights reserved. The n e w e ng l a n d j o u r na l of m e dic i n e Table 2. Secondary Efficacy Outcomes.* Placebo (N=73) Tanezumab, 10 g/kg (N=74) Tanezumab, 25 g/kg (N=75) Tanezumab, 50 g/kg (N=72) Pain subscale -16.22.4 -30.12.3 -36.02.2 Stiffness subscale -16.32.4 -33.52.3 -37.72.2 Physical-function subscale -15.22.3 -30.12.3 Response to therapy according to OMERACT-OARSI criteria by week 16 (% of patients) 43.8 74.3 Outcome Tanezumab, 100 g/kg (N=74) Tanezumab, 200 g/kg (N=72) -29.02.4 -39.62.2 -43.52.3 -34.52.4 -42.72.2 -47.82.4 -34.92.2 -30.82.4 -40.52.2 -43.82.3 84.0 75.0 93.2 93.1 Change in WOMAC score from baseline through week 16 *\tPlus-minus values are means SE. P<0.001 for all comparisons of the five doses of tanezumab with placebo. These analyses were performed on data from the modified intention-to-treat population. \tOne patient who was randomly assigned to receive 25 g of tanezumab per kilogram of body weight instead received 50 g per kilogram. \tAccording to the criteria of the Outcome Measures for Rheumatology Committee and Osteoarthritis Research Society International Standing Committee for Clinical Trials Response Criteria Initiative (OMERACT-OARSI), patients were classified as having had a response if the WOMAC pain or physical-function score decreased by 50% or more and by 20 or more points on the visual-analogue scale or if two of the following three findings were observed: a decrease in the WOMAC pain score by 20% or more and by 10 or more points on the scale, a decrease in the WOMAC physical-function score by 20% or more and by 10 or more points on the scale, or an increase in the score on the patient's global assessment by 20% or more and by 10 or more points on the scale. ezumab groups. Some differences in the onset and duration of these symptoms were noted: allodynia, dysesthesia, and hyperesthesia tended to develop primarily after the first dose of tanezumab had been administered and were relatively short-lived, whereas the onset and duration of paresthesia were more variable. Serious adverse events were reported in 6 patients (2%) receiving tanezumab (appendicitis, bacterial arthritis, cellulitis, spinal stenosis, breast cancer, and syncope) and in 1 patient (1%) receiving placebo (noncardiac chest pain). A total of 6% of tanezumab-treated patients withdrew from the study because of adverse events; no placebotreated patients withdrew because of adverse events (Fig. 1). We observed no clinically important changes in electrocardiographic findings, postural vital signs, or mental status or cognition in any of the study participants, and we did not detect the presence of antitanezumab antibodies in any of the patients assigned to a tanezumab group. One site in the current study also participated in a subsequent phase 3 trial and was closed by the sponsor owing to substantial noncompliance with Good Clinical Practice guidelines and with the protocol in that phase 3 study. Therefore, all analyses from the current study were repeated, with the 23 patients from that site excluded; only very small changes in the efficacy and safety re1528 sults and slight increases in P values were seen when the patients from that site were not included in the analyses (see the Supplementary Appendix, available at NEJM.org). Discussion Two injections of tanezumab a monoclonal antibody that inhibits nerve growth factor 8 weeks apart at doses ranging from 10 to 200 g per kilogram resulted in clinically significant reductions in knee pain, stiffness, and limitations of physical function in patients with moderate-tosevere knee osteoarthritis. Entries in daily pain diaries indicated that differences between tanez umab therapy and placebo were apparent within days after the first injection, and the efficacy persisted throughout the 4-month treatment period. Although the study was not powered to assess dose response, and no formal dose-response analy is was performed, the reductions in pain s appeared to be greater among patients taking higher doses of tanezumab (100 or 200 g per kilogram) than among those taking lower doses, with no clear benefit of the 200-g dose over the 100-g dose. Clinically meaningful pain relief is often described as a reduction in pain intensity of approximately 30% from the baseline level,34,35 and in this study, reductions ranged from 45 to 62% with tanezumab. Furthermore, reductions in n engl j med 363;16 nejm.org october 14, 2010 The New England Journal of Medicine Downloaded from nejm.org at NEW YORK STATE LIB ACQUISITIONS on January 9, 2011. For personal use only. No other uses without permission. Copyright 2010 Massachusetts Medical Society. All rights reserved. Tanezumab for Osteoarthritis of the Knee Table 3. Frequency of Adverse Events. Placebo (N=74) Variable Tanezumab, 10 g/kg (N=74) Tanezumab, 25 g/kg (N=74) Tanezumab, 50 g/kg (N=74) Tanezumab, 100 g/kg (N=74) Tanezumab, 200 g/kg (N=74) number of patients (percent) Any adverse event 41 (55) 51 (69) 49 (66) 44 (59) 51 (69) 58 (78) Treatment-related adverse event 6 (8) 11 (15) 13 (18) 8 (11) 21 (28) 26 (35) Severe adverse event* 2 (3) 6 (8) 3 (4) 3 (4) 3 (4) 3 (4) Treatment-related severe adverse event 0 0 1 (1) 0 1 (1) 0 Serious adverse event 1 (1) 2 (3) 0 2 (3) 0 2 (3) 2 (3) 8 (11) 5 (7) 8 (11) 6 (8) 6 (8) Adverse event occurring in 5% of tanezumabtreated patients Headache Upper respiratory tract infection 4 (5) 2 (3) 6 (8) 5 (7) 7 (9) 7 (9) Arthralgia 0 1 (1) 2 (3) 5 (7) 4 (5) 7 (9) Pain in extremity 0 3 (4) 1 (1) 2 (3) 6 (8) 9 (12) Peripheral edema 2 (3) 0 2 (3) 5 (7) 6 (8) 8 (11) Allodynia 0 0 0 0 1 (1) 1 (1) Burning sensation 1 (1) 0 0 0 1 (1) 0 Adverse event involving abnormal peripheral sensation Dysesthesia 0 0 0 0 1 (1) 1 (1) Hyperesthesia 0 0 0 3 (4) 4 (5) 4 (5) Hypoesthesia 0 1 (1) 6 (8) 2 (3) 5 (7) 5 (7) Neuralgia 0 0 0 0 1 (1) 0 Neuritis 0 0 0 0 1 (1) 0 Pallanesthesia 0 0 1 (1) 0 0 1 (1) Paresthesia 2 (3) 4 (5) 4 (5) 1 (1) 8 (11) 8 (11) Sensory disturbance 0 0 0 1 (1) 1 (1) 2 (3) Sensory loss 0 0 1 (1) 0 0 1 (1) *\tThe severity of an adverse event refers to the maximum intensity of the event. An event was considered to be severe (as compared with mild or moderate) if it interfered substantially with the patient's usual functioning. \tAn adverse event was classified as serious if it was fatal or life-threatening, required or prolonged inpatient hospitalization, was disabling, resulted in a congenital anomaly or birth defect, or required medical or surgical intervention to prevent permanent impairment or damage. No serious adverse event was considered to be treatment-related. \tThe determination of a rate of 5% or higher was made on the basis of all tanezumab groups combined. pain with tanezumab therapy resulted in pain scores that were equal to or lower than those reported by patients at the time of screening, when they were taking their previously prescribed pain medications. The majority of adverse events that were reported by patients taking tanezumab, including abnormal peripheral sensations, were mild to moderate in severity. The occurrence of adverse events appeared to be dose-dependent. Since nerve growth factor is thought to act on small-diameter sensory afferents, the occurrence of paresthesia and other signs associated with large-fiber sensory function is interesting. The more frequent occurrence of these events within a short time after administration of the first dose suggests that there may be transient changes in sensitivity or \"tone\" of different afferent fiber populations, leading to altered sensations. Owing to their largely transient nature, it is unlikely that these adverse events are indicative of neurodegenerative changes. The transient nature of these events is consistent with n engl j med 363;16 nejm.org october 14, 2010 1529 The New England Journal of Medicine Downloaded from nejm.org at NEW YORK STATE LIB ACQUISITIONS on January 9, 2011. For personal use only. No other uses without permission. Copyright 2010 Massachusetts Medical Society. All rights reserved. The n e w e ng l a n d j o u r na l findings in long-term studies of the safety of highdose tanezumab in nonhuman primates.36 Elucidation of the mechanisms underlying these effects and any potential long-term consequences require further investigation. The assessed measures of mental status and cognition were unchanged, suggesting that the effects of tanezumab were limited to the peripheral nervous system. The limitations of this study include the lack of a comparison group receiving a different active treatment, a study population that was too small for a statistical comparison of efficacy according to dose, and short-term exposure to tanezumab (since knee pain from osteoarthritis usually requires long-term treatment). However, our study shows efficacy in patients with more severe osteoarthritis than those in other trials. For example, two studies37,38 of glucosamine and chondroitin in patients with osteoarthritis of the knee excluded patients who had the highest degree of severity on radiography (a Kellgren-Lawrence grade of 4), whereas 17% of our study population had this degree of severity. Baseline scores for knee pain on the visual-analogue scale in these two other studies were 54 and 57, respectively, whereas our population had a mean score for knee pain on the day of