Metachronous Colorectal Neoplasia Risk in Adults Younger Than Age 45 with Advanced Neoplasia at Baseline

Introduction: The risk of metachronous colorectal neoplasia in adults younger than 45 years with advanced lesions at baseline colonoscopy is not well defined. Identifying non-endoscopic factors associated with the risk of neoplasia on surveillance colonoscopy may help advise patients on risk mitigation strategies. The aims of the study are: 1) To examine risk for metachronous neoplasia in adults < 45 years of age who had advanced neoplasia (AN) at baseline colonoscopy, 2) To determine factors associated with metachronous neoplasia in these patients. Methods: Individuals 18-44 years of age with ≥ 1 AN [tubular adenoma (TA) >10 mm or with villous features or high-grade dysplasia (HGD), sessile serrated lesion (SSL) >10 mm or with dysplasia, traditional serrated adenoma (TSA) or invasive adenocarcinoma] who had colonoscopy > 6 months after their baseline colonoscopy were included. Patients with inflammatory bowel disease or inherited CRC syndromes were excluded. Age, gender, race, ethnicity, BMI, alcohol and tobacco use, diabetes (ICD codes, or Hba1c >6.4%), hyperlipidemia (ICD codes or LDL > 130 mg/dl), vitamin D level, family history of CRC, and colonoscopy indications were obtained. Individuals were divided into 3 groups based on metachronous colonoscopy findings: 1) normal (no polyps or hyperplastic polyps), 2) non advanced neoplasia (NAN) [TA < 10mm, SSL < 10 mm without dysplasia] and 3) AN. Multivariable logistic regression model were performed. Results: Of 651 patients included, 331 (50.8%) had ≥ 1 follow up colonoscopies during a mean follow up of 43 months. Follow up findings were normal in 194 (58.6%), NAN in 109 (32.9%) and AN in 28 (8.4%). No adenocarcinoma was diagnosed on follow up. Patients with metachronous NAN and AN had longer follow-up time (50.3 vs 45.3 vs 38.7 months respectively, P< 0.001) and more surveillance colonoscopies (1.1 vs 1.7 vs 1.6 respectively, P< 0.001) compared to patients with a normal colonoscopy. In the multivariable model comparing predictors of a composite endpoint (AN and NAN), only BMI was found to be statistically significant OR 1.072 (P=0.016) (Table 1b). Conclusion: In our cohort, we observed AN in < 10% of adults < 45 years of age during a follow up approximating 4 years. Except for BMI, no clinical or pathological factors were associated with metachronous neoplasia. Personalized surveillance colonoscopy and providing weight management in young patients with high BMI may help reduce risk in these young patients. Table 1. - (A) Baseline Characteristics by Type of Metachronous Neoplasia B) Results of Logistic Regression Model Predicting Metachronous Neoplasia (NAN and AN) (A) Baseline Factors Total (N=331) Normal (N=194) NAN (N=109) AA (N=28) P-value Gender* 0.13c Male 160(48.5) 85(43.8) 60(55.6) 15(53.6) Female 170(51.5) 109 (56.2) 48(44.4) 13(46.4) Age at baseline colonoscopy* 39.0±4.8 38.6±4.9 39.4±4.2 39.3±5.7 0.35a Race* 0.33c Black 52 (15.8) 35(18.0) 14(13.0) 3(10.7) White 246(74.5) 144(74.2) 82(75.9) 20(71.4) Other 32(9.7) 15(7.7) 12(11.1) 5(17.9) Ethnic Group* 0.16c Hispanic 22(6.7) 10(5.2) 8(7.5) 4(14.8) Non-Hispanic 304(93.3) 183(94.8) 98(92.5) 23(85.2) BMI at baseline colonoscopy* 29.5±6.9 28.8±6.6 30.7±7.4 29.2±6.4 0.066a Hyperlipidemia 0.47c No 282(85.2) 163(84.0) 93(85.3) 26(92.9) Yes 49(14.8) 31(16.0) 16(14.7) 2(7.1) Vitamin D level* 27.9±14.5 28.1±13.4 26.2±13.2 33.2±22.1 0.25a Diabetes 0.33c No 315(95.2) 187(96.4) 101(92.7) 27(96.4) Yes 16(4.8) 7(3.6) 8(7.3) 1(3.6) Aspirin Use* 0.95c No 137(77.4) 82(78.1) 41(75.9) 14(77.8) Yes 40(22.6) 23(21.9) 13(24.1) 4(22.2) Baseline colonoscopy AN lesion subtype 0.56c Invasive adenocarcinoma 21(6.3) 13(6.7) 7(6.4) 1(3.6) TVA or TA with HGD 30(9.1) 18(9.3) 10(9.2) 2(7.1) TVA 97(29.3) 54(27.8) 31(28.4) 12(42.9) TA > 10 mm 140(42.3) 89(45.9) 41(37.6) 10(35.7) SSP >10 mm 35(10.6) 15(7.7) 17(15.6) 3(10.7) TSA 8(2.4) 5(2.6) 3(2.8) 0(0.00) Tobacco Use* 0.33c Never 175(53.0) 108(55.7) 53(49.1) 14(50.0) Quit 90(27.3) 50(25.8) 29(26.9) 11(39.3) Yes 65(19.7) 36(18.6) 26(24.1) 3(10.7) Alcohol Use* 0.56c No 96(30.5) 58(31.2) 31(30.4) 7(25.9) Not Currently 27(8.6) 14(7.5) 12(11.8) 1(3.7) Yes 192(61.0) 114(61.3) 59(57.8) 19(70.4) Family History of CRC* 0.72c No 196(59.4) 120(61.9) 62(57.4) 14(50.0) Unknown 2(0.61) 1(0.52) 1(0.93) 0(0.00) Yes 132(40.0) 73(37.6) 45(41.7) 14(50.0) Follow up Time (months) 43.0±26.2 38.7±23.32 50.3±29.01 45.3±28.3 < 0.001a Number of follow up colonoscopies 1.3±0.72 1.1±0.3823 1.7±0.921 1.6±0.951 < 0.001a (B) Full Model 95% Wald CI 95% Wald CI Factors Contrast Odds Ratio Lower Upper P-value Gender Male vs Female 1.222 0.474 3.15 0.678 Hyperlipidemia Yes vs No 0.511 0.135 1.93 0.322 Diabetes Yes vs No 1.743 0.364 8.354 0.487 Aspirin Use Yes vs No 1.03 0.357 2.972 0.957 Family History of CRC Yes vs No 1.009 0.396 2.573 0.984 BMI 1.059 0.992 1.13 0.085 Age at baseline 1.066 0.954 1.19 0.258 Vitamin D 1.001 0.97 1.032 0.969 Reduced Model 95% Wald CI 95% Wald CI Variable Odds Ratio Lower Upper P-value BMI 1.072 1.013 1.134 0.016 *Data not available for all subjects. Missing values: Female = 1, Race = 1, Ethnic Group = 5, Family History = 1, ASA = 154, Tobacco Use = 1, Alcohol Use = 16. Statistics presented as Mean ± SD, Median [P25, P75], Median (min, max) or N (column %). P-values: a=ANOVA, b=Kruskal-Wallis test, c=Pearson's chi-square test, d=Fisher's Exact test. 1: Significantly different from Normal 2: Significantly different from NAN 3: Significantly different from AN.