Colonoscopy in Nonagenarians Is Safe and May Be Associated With Clinical Benefit

Asher Shafrir, MD; Benjamin Koslowsky, MD; Dov Wengrower, MD; Eran Goldin, MD; Dan M. Livovsky, MD


J Am Geriatr Soc. 2019;67(6):1158-1163. 

In This Article


Characteristics of the Population

A total of 346 colonoscopies were studied (128 nonagenarians and 218 octogenarians). The mean ages were 83.3 ± 2.4 and 92.2 ± 2.8 years. Nonagenarians were more likely to undergo a colonoscopy while hospitalized (56.2 vs 23.4%; P < .001), were more likely to undergo the examination due to rectal bleeding or volvulus (35.2 vs 25.2 and 10.9 vs 0.5%, respectively; P < .001) and less likely due to postpolypectomy surveillance or constipation (11.7 vs 25.7 and 0 vs 6.9%, respectively; P < .001) (Table 1).

Preparation regimens did not differ among groups, but nonagenarians were significantly more likely to have a very poor preparation (17.2 vs 9.2%; P = .03).

Completion rates including all procedures was lower in nonagenarians (78.9% vs 88.1; P = .03). However, when excluding procedures that were terminated due to poor preparation, completion rates were comparable (94.5 vs 91.4%). Overall, severe adverse events were nonsignificantly different between groups, 0.9% in octogenarians vs 1.6% in nonagenarians (P = .6).

Short-term Prognosis

The 30-day mortality in the entire cohort was 1.7% (n = 6); 3.9% (five patients) in nonagenarians and 0.4% (one) in octogenarians (P = .027). Four patients (three nonagenarians and one 83-year-old patient) died within 1 week after being diagnosed with adenocarcinoma. Three had a large obstructing tumor and were operated on urgently; following surgery, they suffered multiorgan failure. One 91-year-old patient had metastatic CRC involving the brain at the time of diagnosis and died shortly after. One patient died from overwhelming sepsis secondary to urinary tract infection 2 weeks after the initial colonoscopy. And the sixth patient died from uncontrollable bleeding originating from a large angioectasia in the terminal ileum.

Number of Procedures Performed on Nonagenarians

The absolute and percentage of colonoscopies performed on nonagenarians during a 4-year observation showed a significant increase from 0.34% in 2013 to 0.62% in 2016, representing a relative increase of 80% (Supplementary Figure 1).

Supplementary Figure S1.

Increasing numbers (a) and frequency (b) of colonoscopies in nonagenarians in our medical center

Diagnostic Yield

Advanced adenomas and carcinoma were more common in nonagenarians (25.8 vs 16.5%, P = .03 and 14.8 vs 6.4%, P = .01, respectively). Nonagenarians had larger polyps (18.8 (range = 1-60) vs 12.9 (5-60) mm; P = .012). In contrast, nonadvanced adenomas were more common in octogenarians (Figure 1A and B). Polyps in octogenarians were more likely to be adenoma with low-grade dysplasia, but this did not reach statistical significance (Table 2).

Figure 1.

Distribution of any neoplasia found by age group (A) and distribution of neoplasia found by severity (nonadvanced adenoma, advanced adenoma, or colorectal cancer [CRC]) by age group (B)

In univariate analysis, increasing age, inpatient status, history of previous polypectomy, and anemia were associated with higher odds of carcinoma. Remarkably, no tumors were found in patients who underwent colonoscopy for constipation or diarrhea. In the multivariate logistic regression model, only increasing age was associated with a trend for increased OR [1.072 (0.99-1.15); P = .075] of carcinoma (Supplementary Table S1).

Follow-up and Management

Data regarding follow-up and management was available for 31 of 34 patients diagnosed with CRC; two octogenarians and one nonagenarian were lost to follow-up. All octogenarians (n = 12) diagnosed with adenocarcinoma underwent surgery compared with 52.6% (10/19) of nonagenarians (P = .01). The median follow-up time until mortality or censor was significantly longer for octogenarians (1006 [IQR = 438.5-1566]) vs 200.5 days ([55.2-854]; P < .019). Overall mortality related to CRC was higher among nonagenarians (73 vs 25%; P = .023). Mean survival including all nonagenarians who underwent surgery was 605 days (IQR = 11-878) compared with 112 days (48-341) in those treated conservatively (P = .7), nevertheless, if the patients who died in the immediate postoperative period are excluded, the median survival was 878 (IQR = 741-1110; P = .13) (Supplementary Table S2). On survival analysis, significantly longer survival was observed in octogenarians (Supplementary Figure S2; log-rank test P = .001) and in patients who underwent surgery irrespective of age (Figure 2; log-rank test P < .001). Remarkably, nonagenarians with CRC who underwent surgery (including those who had emergency surgery with postoperative mortality) had a trend for longer survival (Figure 3; log-rank test P = .055).

Figure 2.

Kaplan-Meier survival curve of all patients who were diagnosed with colorectal carcinoma (CRC) and were treated either with surgery or with palliation only

Figure 3.

Kaplan-Meier survival curve comparing the survival of nonagenarians with colorectal carcinoma and who were treated either with surgery or palliation alone

Supplementary Figure S2.

Kaplan Meier survival curve comparing the survival of nonagenarians and octogenarians patients who were diagnosed with colorectal carcinoma.