Figure 3.13 Curves of log[9log S(t)] for the two hypertension groups. Table 3.9 Significant Variables (at 0.05 Level) Identified by Proportional Hazards Model Relative Risk@ Ratio Regression p Value of Variable? Coefficient (Ward Test) Favorable Unfavorable Risk Age 0.0558 :0.001 9.32 28.45 3.05 Hypertension: 0.6360 :0.001 1.00 1.89 1.89 1, yes, 0, no Duration of 0.0559 :0.001 1.32 2.19 1.66 diabetes Fasting plasma 0.0023 :0.010 1.35 1.58 1.17 glucose BMI 90.0330 0.035 0.32 0.44 0.72 Proteinuria: 0.3744 0.025 1.00 1.45 1.45 1, yes, 0, no Use of diuretics: 0.4191 0.030 1.00 1.52 1.52 1, yes; 0, no ? Variables are listed in order of entry into model with a p-value limit for entry of 0.05. @ Favorable categories are 40 years of age, no hypertension, duration of diabetes 5 years, fasting plasma glucose 130mg/dL, BMI 35, no proteinuria, and no diuretics use. Unfavorable categories are 60 years of age, hypertensive, duration of diabetes 14 years, fasting plasma glucose 200 mg/dL, BMI 25, having proteinuria, and diuretics use. EXAMPLE 3.5: IDENTIFICATION OF RISK FACTORS 41 developed the eye disease during the 10 to 16-year follow-up period (average follow-up time 12.7 years). Twelve potential factors (assessed at time of baseline examination) were examined by univariate and multivariate methods for their relationship to retinopathy (RET): age, gender, duration of diabetes (DUR), fasting plasma glucose (GLU), initial treatment (TRT), systolic (SBP) and diastolic blood pressure (DBP), body mass index (BMI), plasma cholesterol (TC), plasma triglyceride (TG), and presence of macrovascular disease (LVD) or renal disease (RD). Table 3.10 gives the data for the first 40 patients. Among other things, the authors related these variables to the development of retinopathy. 1. Examine the individual relationship of each variable to the development of diabetic retinopathy. Table 3.11 gives some summary statistics of the eight continuous variables for patients who have developed retinopathy and for those who have not. Notice that patients who have developed the disease were younger at baseline and had much higher fasting plasma glucose, systolic and diastolic blood pressure, and plasma triglyceride than did patients who have not. Table 3.12 summarizes the contingency table analysis of retinopathy incidence rates. The number of patients at risk of developing retinopathy and the number of patients who developed the disease (and rate) are given by subcategory of each potential risk factor. Using the chi-square test, it is found that there was a significant difference in the retinopathy rate among the subcategories of several variables using a significance level of 0.05: duration of diabetes, fasting plasma glucose, systolic and diastolic blood pressure, and treatment. It appears that patients with poor glucose control or high blood pressure or treated with oral agents or insulin have a higher incidence of retinopathy. In addition, patients with high triglyceride levels tend to have higher incidence of retinopathy (p : 0.064). However, patients who had developed macrovascular disease at the time of baseline examination had a lower retinopathy incidence. The authors state that this may be due to the fact that 68% of the patients who had macrovascular disease either died (54%) during the follow-up period or were lost to follow-up (14%). Many of these patients may have developed retinopathy, particularly the patients who have died, but were not included. Therefore, the lower incidence of retinopathy in patients who had macrovascular disease at baseline is probably the result of a selection bias. Similarly, the large number of death plus the losses to follow-up may also contribute to the drop in retinopathy rate in patients who had had diabetes for more than 12 years at baseline. Among the 80 patients in this duration of diabetes category, 56% have died and 10% did not participate in the follow-up examination. The large number of deaths may also be responsible for the finding that patients who survived long enough to develop retinopathy were younger at baseline. The deceased patients were significantly older (mean 57 years) than the survivors who participated in the follow-up examination (mean 48 years). 42 EXAMPLES OF SURVIVAL DATA ANALYSIS Table 3.10 First 40 Patients Involved in Study of Risk Factors in Development of Diabetic Retinopathy Patient RET? Age Gender DUR TRT@ GLU SBP DBP TC TG BMI LVD? RD? 1 1 47.6 M 4 2 100 156 98 195 405 33.1 0 0 2 0 54.4 M 7 1 112 112 78 204 77 31.7 0 0 3 0 50.8 M 4 2 83 134 80 206 178 41.5 1 0 4 1 49.4 F 5 2 276 102 70 190 222 24.1 0 0 5 1 50.0 M 2 2 104 142 86 178 100 39.5 0 0 6 1 50.7 F 7 2 242 142 78 217 268 31.6 0 1 7 1 35.3 F 2 2 130 134 80 390 564 47.0 0 0 8 0 50.2 F 5 2 130 100 70 174 128 29.8 0 0 9 1 45.0 M 5 2 115 134 100 238 177 18.9 1 0 10 0 38.3 M 2 2 110 132 80 204 180 31.2 1 0 11 0 45.8 F 5 3 130 118 68 185 316 26.6 1 0 12 1 51.6 F 3 2 141 112 78 152 77 31.2 0 1 13 1 36.3 M 6 2 238 142 94 194 162 24.2 0 0 14 1 44.7 F 16 2 190 152 90 132 161 32.0 0 0 15 1 37.2 F 1 2 126 136 90 133 211 32.7 0 0 16 1 52.6 F 9 2 159 140 76 151 132 26.7 0 0 17 1 44.1 M 1 1 116 126 76 251 153 33.3 0 0 18 1 35.9 M 3 1 120 132 80 129 76 30.1 0 0 19 1 50.4 M 1 2 128 144 90 190 123 27.7 0 0 20 1 48.0 M 1 2 95 128 74 207 59 28.1 1 0 21 0 47.5 F 1 2 85 124 82 161 190 31.6 1 0 22 1 50.1 F 6 2 138 106 72 181 135 30.5 0 1 (Continued overleaf ) 43 Table 3.10 Continued Patient RET? Age Gender DUR TRT@ GLU SBP DBP TC TG BMI LVD? RD? 23 0 43.3 F 1 1 104 128 86 204 198 26.1 0 0 24 1 54.5 M 0 3 104 142 84 490 540 30.8 1 0 25 1 52.2 F 3 3 304 132 84 192 119 36.9 1 0 26 1 53.3 F 3 2 249 128 72 120 85 35.5 0 0 27 1 64.3 F 6 2 297 138 80 145 64 30.1 1 0 28 1 44.6 F 4 1 139 112 80 156 111 36.1 0 1 29 1 47.1 F 2 2 169 130 84 198 99 39.0 0 1 30 1 46.5 F 7 1 159 128 78 238 157 34.5 0 0 31 0 51.5 F 3 1 147 128 78 185 182 32.0 0 0 32 1 59.5 M 3 2 180 132 78 188 308 28.1 1 0 33 1 52.0 F 6 2 183 142 84 175 68 37.6 0 0 34 1 45.7 F 6 2 180 138 80 179 189 44.1 1 0 35 1 48.2 F 18 2 267 158 100 195 112 25.2 1 0 36 1 57.4 M 0 1 159 172 108 219 294 33.7 0 1 37 1 42.0 F 4 2 158 106 68 224 157 33.6 0 0 38 1 50.7 F 1 2 211 142 84 390 645 37.1 0 0 39 1 53.8 F 0 1 177 154 80 175 208 35.3 0 1 40 0 56.9 M 1 1 98 116 70 146 97 26.5 0 1 ? 1, yes; 0, no. @ 1, diet only; 2, oral agent; 3, insulin. 44 Table 3.11 Summary Statistics for Eight Variables by Retinopathy Status at Follow-up Retinopathy Status No Yes Variable Mean S.D. Mean S.D. p Value Age 50.0 9.0 47.2 7.4 0.01 Duration of diabetes 4.2 4.5 4.8 4.4 0.34 Fasting plasma glucose 141.8 65.6 196.3 76.6 :0.0001 Systolic blood pressure 128.0 15.7 132.6 17.3 0.04 Diastolic blood pressure 80.3 10.8 84.9 10.1 :0.001 Body mass index 32.3 6.3 32.5 5.9 0.76 Cholesterol 204.4 66.0 206.8 58.7 0.76 Triglyceride 180.5 111.1 234.4 273.3 0.01 2. Examine the simultaneous relationship of the variables to the development of retinopathy. Univariate analysis of each variable using the contingency table or the chi-square test gives a preliminary idea of which individual variable might be of prognostic importance. The simultaneous effect of all the variables can be analyzed by the linear logistic regression model (discussed in Section 14.2) to determine the relative importance of each. The 12 variables were fitted to the linear logistic regression model using a stepwise selection procedure. The variables most significantly related to the development of retinopathy were found to be initial treatment, fasting plasma glucose, age, and diastolic blood pressure (p - 0.001). Table 3.13 gives the regression coefficients of the four most significant variables (p - 0.05), the standard errors, and adjusted odds ratios [exp(coefficient)]. The p values used here are the significance levels based on the likelihood ratio test or the improvement in the maximum likelihood due to the addition of the variable in the stepwise procedure. This method is more powerful than the Wald test, which is based on the standardized regression coefficients (Chapter 14). The results are consistent with those in the univariate analysis. On the basis of the regression coefficients, the probability of developing retinopathy during a 10 to 16-year follow-up can be estimated by substituting values of the risk factors into the regression equation, log P 1 9 P :92.373 ; 1.495 (oral agent) ; 0.882 (insulin) ; 0.014 (GLU) 9 0.074 (age) ; 0.048 (DBP) For example, for a 50-year-old patient who is on oral agents and whose fasting plasma glucose and diastolic blood pressure are 170 mg/dl and 95 mmHg, EXAMPLE 3.5: IDENTIFICATION OF RISK FACTORS 45 Table 3.12 Cumulative Incidence Rates of Retinopathy by Baseline Variables Developed Retinopathy Number of Variable Persons at Risk Number Percent p value Gender Female 211 151 71.6 0.384 Male 101 77 76.2 Age (yr) :35 13 10 76.9 35—44 101 77 76.2 0.242 45—54 155 115 74.2 .55 43 26 60.5 Duration of diabetes (yr) :4 153 105 68.6 4—7 113 86 76.1 0.033 8—11 23 22 95.7 .12 23 15 65.2 Fasting plama glucose (mg/dl) :140 117 62 53.0 140—199 90 74 82.2 :0.001 .200 105 92 87.6 Systolic blood pressure (mmHg) :130 145 95 65.5 130—159 149 115 78.8 0.016 .160 20 18 85.7 Diastolic blood pressure (mmHg) :85 179 118 65.9 85—94 87 73 83.9 0.004 .95 46 37 80.4 Plasma cholesterol (mg/dl) :240 267 193 72.3 0.442 .240 45 35 77.8 Plasma triglyceride (mg/dl) :250 237 167 70.5 0.064 .250 75 61 81.3 Body mass index (kg/m) :28 73 49 67.1 28—33 121 94 77.7 0.261 .34 118 85 72.0 Renal disease No 251 179 71.3 0.155 Yes 61 49 80.3 Macrovascular disease No 205 157 76.6 0.053 Yes 107 71 66.4 Treatment (initial) Diet alone 115 62 53.9 Oral agent 158 136 86.1 :0.001 Insulin 37 29 78.4 46 EXAMPLES OF SURVIVAL DATA ANALYSIS Table 3.13 Results of Logistic Regression Analysis Standard Variable Coefficient Error exp(coefficient) Coefficient/S.E. Constant 92.373 1.557 Initial treatment Oral agent 1.495 0.330 4.459 4.53 Insulin 0.882 0.488 2.416 1.81 Fasting plasma 0.014 0003 1.014 4.67glucose Age 90.074 0019 0.929 93.89 Diastolic blood 0.048 0.015 1.049 3.20pressure respectively, the chance of developing retinopathy in the next 10 to 16 years is 91%. The linear logistic regression model is useful in identifying important risk factors. However, complete measurements of all the variables are needed; missing data are a problem. In this example, complete data are available on most of the patients. This may not always be the case. Although there are methods of coping with missing data (discussed in Section 11.1), none is perfect. Thus it is extremely important for investigators to make every effort to obtain complete data on every subject. Bibliographical Remarks It is impossible to cite all the published examples of survival data analysis similar to those in this chapter. Other similar studies can be found in the literature: for example, Biometrics, Biometrika, Cancer, Journal of Chronic Disease, Journal of the National Cancer Institute, American Journal of Epi- demiology, Journal of the American Medical Association, and New England Journal of Medicine. An easy way to find examples is to use the National Library of Medicine’s Web site and search the file PubMed with appropriate keywords. EXERCISES The four sets of data below are taken from actual research situations. Although the data can be used for various analyses throughout the book, the reader is asked here only to describe in detail how the data can be analyzed. The data appear in examples and other exercises in subsequent chapters. 3.1 Thirty-three patients with hypernephroma were treated with combined chemotherapy, immunotherapy, and hormonal therapy. Exercise Table 3.1 EXERCISES 47 Exercise Table 3.1 Data for 33 Patients with Hypernephroma Date of Date Death or Skin Test ResultsA Treatment Last Patient Age Gender Started Response? Follow-up Status@ Monilia Mumps PPD PHA SK-SD 1 53 F 3/31/77 1 10/1/77 0 7;723;23 0;025;25 0;0 2 61 M 6/18/76 0 8/21/76 1 10;10 15;20 0;013;13 9;9 3 53 F 2/1/77 3 10/1/77 0 0;07;70;025;25 0;0 4 48 M 12/19/74 2 1/15/76 1 0;00;00;00;00;0 5 55 M 11/10/75 0 1/15/76 1 12;12 ND 10;10 8;85;5 6 62 F 10/7/74 2 4/5/75 1 10;10 5;50;07;75;5 7 57 M 10/28/74 0 1/6/75 1 15;15 15;15 0;00;010;10 8 53 M 10/6/75 2 6/18/77 1 0;0ND0;012;12 0;0 9 45 M 4/11/77 0 10/1/77 0 6;44;40;00;00;0 10 58 M 8/4/76 3 2/11/77 1 13;13 13;13 22;22 23;23 0;0 11 61 F 1/1/77 3 10/1/77 0 0;08;817;17 11;11 0;0 12 61 M 7/25/76 1 10/1/77 0 9;912;12 0;020;20 0;0 13 77 M 5/8/75 0 9/26/75 1 0;00;00;00;00;0 14 55 M 4/27/77 2 10/1/77 0 0;00;015;15 10;10 0;0 15 50 M 4/20/77 3 10/1/77 0 0;014;14 5;532;32 21;21 16 42 M 8/24/76 0 10/1/77 0 11;11 7;70;012;12 0;0 48 17 50 F 1/8/75 0 6/30/75 1 0;00;00;00;00;0 18 66 F 9/8/76 3 10/1/77 0 9;910;10 6;615;15 11;11 19 58 M 2/18/75 0 10/1/77 0 0;00;00;00;0ND 20 62 M 5/12/76 0 10/17/76 1 2;2NDND3;32;2 21 71 F 10/22/76 3 12/12/76 1 10;10 6;60;012;12 0;0 22 44 M 6/6/77 3 10/1/77 0 10;10 10;10 0;020;20 0;0 23 69 M 6/21/76 0 10/13/76 1 0;015;15 25;25 25;25 0;0 24 56 M 6/7/77 2 10/1/77 0 0;07;70;00;00;0 25 57 M 11/16/76 0 12/10/76 1 11;11 5;50;020;20 0;0 26 69 M 5/10/77 0 7/25/77 1 0;00;00;015;15 0;0 27 60 M 6/29/77 0 7/7/77 1 0;00;00;026;26 0;0 28 60 M 7/21/75 3 10/1/77 0 11;11 20;20 10;10 18;18 0;0 29 72 M 7/19/75 0 10/18/75 1 10;10 0;07;710;10 0;0 30 42 F 3/3/75 0 4/23/75 1 0;0ND0;00;00;0 31 57 M 2/24/77 2 10/1/77 0 5;58;80;025;15 0;0 32 66 M 6/15/77 3 10/1/77 0 0;015;15 0;010;10 0;0 33 59 M 3/4/77 0 4/2/77 1 0;00;00;016;16 0;0 Source: Data courtesy of Richard Ishmael. ? 0, no response; 1, complete response; 2, partial response; 3, stable. @ 0, alive; 1, dead. A ND, not done. 49 gives the age, gender, date treatment began, response status, date of death or last follow-up, survival status, and results of five pretreatment skin tests. The investigator is interested in the response and survival of the patients and in identifying prognostic factors. How would you analyze the data? 3.2 In a study undertaken to compare the treatments given to hyperneph- roma patients and to relate response and survival to surgery, metastasis, and treatment time, data from 58 patients were collected (Exercise Table 3.2). How would you analyze the data to answer these questions? (a) Do patients who had nephrectomy have a higher response rate? (b) Is the time of nephrectomy related to response and survival? (c) Are there significant differences between the treatments? (d) What are the most important variables related to response and survival? 3.3 Exercise Table 3.3 gives the age, gender, family history of melanoma, remission duration, survival time, stage, and results of six pretreatment skin tests (the larger diameter is given) of 102 stage 3 and 4 melanoma patients (Lee et al., 1982). (a) Study the immunocompetence of melanoma patients by investigating skin test results. (b) Determine if age, gender, or pretreatment skin test results are predic- tive to remission and survival time. (c) Find theoretical distributions that describe the survival and remission patterns. 3.4 One hundred and forty-nine diabetic patients were followed for 17 years (a subset of data from Lee et al., 1988). Exercise Table 3.4 gives the survival time from baseline examination, survival status, and several potential prognostic factors at baseline: age, body mass index (BMI), age at diagnosis of diabetes, smoking status, systolic blood pressure (SBP), diastolic blood pressure (DBP), electrocardiogram reading (ECG), and whether the patient had any coronary heart disease (CHD). Identify the important prognostic factors that are associated with survival. 50 EXAMPLES OF SURVIVAL DATA ANALYSIS [...]... 52 41 49 44 37 51 47 45 38 35 50 53 48 40 43 54 52 69 32. 3 34.5 18.9 32. 0 33.9 23 .7 24 .8 26 .6 39 .2 32. 7 33.5 32. 2 24 .2 31.6 30.7 28 .0 32. 0 32. 7 24 .2 18.7 25 .6 22 .8 30.1 27 .7 27 .6 28 .1 31.7 26 .1 30.8 36.9 24 .2 52 47 40 31 30 28 43 41 35 36 43 49 30 48 39 35 29 36 42 42 36 27 33 49 49 47 37 42 54 50 63 1 2 1 1 2 0 0 2 2 2 1 2 2 1 2 2 0 2 2 0 0 2 0 1 2 1 2 2 1 1 1 1 32 150 134 1 42 124 1 02 134 118 1 92 122 ... 25 .2 25.3 25 .8 33.7 39.5 32. 9 37.1 35.3 29 .3 22 .1 23 .6 26 .1 32. 5 29 .8 24 .4 26 .3 30.8 29 .4 81 39 42 60 43 41 54 45 28 40 59 54 30 50 34 57 55 48 50 53 70 33 43 55 45 67 80 54 46 46 0 0 0 2 0 2 1 0 1 0 1 0 2 0 1 2 1 0 0 2 0 2 0 2 2 0 1 0 2 0 1 42 160 122 1 62 1 32 116 1 52 144 98 138 138 184 158 176 118 1 72 1 82 144 1 42 154 122 22 2 150 1 42 128 122 1 62 1 72 1 32 1 12 88 78 68 98 72 60 84 68 68 76 78 80 98 96 72. .. : 0.046 1 1 1 ; ; 29 ;30 28 ;29 27 ;28 : 0.033 1 1 ; 29 ;30 28 ;29 1 29 ;30 Standard Error of S(t) Table 4.3 Calculation of S(t) and Standard Error of S(t) for 30 Rats on a Low-Fat Diet in Table 3.4 : 0.055 73 153 177 181 191 20 0; 20 0; 20 0; 20 0; 20 0; 20 0; 20 0; 20 0; 20 0; 20 0; 20 0; 20 0; 20 0; 20 0; 6 2 7 8 17 18 19 20 21 22 23 24 25 26 27 28 29 30 13 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15... 48 49 50 51 52 53 54 55 56 57 58 70 6 8 12 20 8 999 12 181 20 14 26 16 30 20 0 1 1 1 1 1 0 1 0 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1 1 1 0 1 1 1 0 1 1 0 0 0 1 0 0 1 0 0 0 0 0 54 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Patient 58 50 76 66 33 55 25 23 30 34 34 26 27 72 70 82 43 52 34 48 62 49 46 53 21 25 35 Age 2 2 1 2 1 2 2 1 1 1 1 2 1 2 2 2 1 1 1 1 1 1 1 2 2 1 2 Gender 9 0 0... 57 58 51 33 36 52 64 31 69 59 38 49 49 68 40 36 60 74 61 54 Age (yr) 29 .1 30.1 31.0 34.0 38.1 37.0 31 .2 38.8 22 .3 25 .0 31.3 59.7 34.0 29 .4 43 .2 35.1 37.0 27 .1 27 .6 25 .2 BMI 47 56 37 33 33 46 57 29 56 58 38 49 41 66 41 32 54 54 51 51 Age at Diagnosis (yr) 1 0 2 2 1 0 2 1 0 0 2 1 0 1 1 2 0 1 0 0 Smoking Status@ 138 128 1 32 120 122 140 1 72 136 1 52 126 104 1 42 128 122 122 122 122 168 1 62 116 SBP (mmHg)... 127 128 129 130 131 1 32 133 134 135 136 137 138 139 140 141 1 42 143 144 145 146 147 148 149 35 46 40 53 66 61 41 64 41 46 80 63 72 41 52 53 61 53 75 40 61 62 49 25 .8 32. 2 41.6 39.8 26 .6 33.3 27 .7 26 .6 25 .0 54.3 29 .4 33.1 27 .3 36.9 40 .2 32. 7 33 .2 41.4 35.8 34.0 19.9 30.6 30.8 34 42 41 52 54 55 38 51 38 45 79 60 68 33 36 48 57 47 66 38 37 49 47 2 2 2 0 1 0 1 2 2 1 1 1 1 0 0 2 1 1 0 2 0 0 1 126 180 1 32. .. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Patient 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 1 1 1 1 1 0 1 Status? 12. 4 12. 4 9.6 7 .2 14.1 14.1 12. 4 14 .2 12. 4 14.5 12. 4 10.8 10.9 10.3 13.6 11.9 12. 5 5.9 12. 4 14.1 9.8 7 .2 Survival Time (yr) 44 49 49 47 43 47 50 36 50 49 50 54 42 44 40 48 50 47 38 35 51 40 Age (yr) 34 .2 32. 6 22 .0 37.9 42. 2 33.1 36.5 38.5 41.5 34.1 39.5 42. 9 29 .8 33 .2 27.5 25 .3 31.6 26 .3... 26 .3 32. 4 47.0 26 .5 43.9 BMI Exercise Table 3.4 Data of 149 Diabetic Patients 41 48 35 45 42 44 48 33 47 45 48 43 36 43 26 48 44 38 36 33 47 34 Age at Diagnosis (yr) 0 2 2 0 2 0 0 2 1 0 2 0 2 2 2 0 1 1 2 1 2 0 Smoking Status@ 1 32 130 108 128 1 42 156 140 144 134 1 02 1 42 128 156 1 02 146 120 1 42 144 150 134 130 122 SBP (mmHg) Variable at Baseline 96 72 58 76 80 94 86 88 78 68 84 74 86 58 98 68 76 82 98... 118 1 92 122 122 1 12 1 42 1 52 1 12 118 1 52 136 134 130 108 126 1 32 144 126 128 1 32 128 1 42 1 32 148 80 88 98 90 66 60 80 66 108 78 92 74 90 96 74 84 88 88 90 78 72 66 78 88 68 70 82 80 80 80 78 0 1 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 1 1 0 (Continued overleaf ) 1 3 1 2 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 2 1 2 1 1 1 1 1 1 2 2 1 60 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75... 1 1 1 1 1 1 Status? 16.9 3.6 10 .2 15.7 12. 0 6.7 11.6 2. 0 10 .2 3.6 15.4 11.3 10.3 5.8 8.0 14.6 11.4 7 .2 5.5 11.1 16.5 10.9 2. 5 Survival Time (yr) Exercise Table 3.4 Continued 38 50 64 44 38 62 47 78 49 63 71 51 59 50 66 42 40 67 86 52 42 60 75 Age (yr) 27 .5 27 .3 30.1 36.1 43.1 34.6 39.0 28 .7 28 .2 25.1 26 .0 32. 0 28 .1 26 .1 45.3 30.0 35.7 28 .1 32. 9 37.6 43.4 25 .4 49.7 BMI 26 44 58 41 39 58 45 77 43 46 59 . 66 2 1 31 1 12. 4 40 39 .2 35 2 1 92 108 1 0 32 1 14.4 44 32. 7 36 2 122 78 1 0 33 1 14 .2 48 33.5 43 1 122 92 1 0 34 1 14.5 51 32. 2 49 2 1 12 74 1 0 35 1 12. 4 36 24 .2 30 2 1 42 90 1 0 36 1 14.3 52 31.6. 32. 4 36 2 150 98 2 1 20 1 14.1 35 47.0 33 1 134 78 1 0 21 0 9.8 51 26 .5 47 2 130 76 1 0 22 1 7 .2 40 43.9 34 0 122 92 1 0 58 23 1 3.5 54 32. 3 52 1 1 32 80 1 0 24 1 0.0 53 34.5 47 2 150 88 3 1 25 . 58 M 2/ 18/75 0 10/1/77 0 0;00;00;00;0ND 20 62 M 5/ 12/ 76 0 10/17/76 1 2; 2NDND3; 32; 2 21 71 F 10 /22 /76 3 12/ 12/ 76 1 10;10 6;60;0 12; 12 0;0 22 44 M 6/6/77 3 10/1/77 0 10;10 10;10 0; 020 ;20 0;0 23 69