Serious gastrointestinal colonoscopy complications

Many studies have reported the risks of serious colonoscopy complications, and most commonly have tracked the risks of colonic perforation or gastrointestinal bleeding. Other serious complications include postpolypectomy syndrome and diverticulitis. It is known that serious complications, in particular gastrointestinal bleeding, can occur at delayed intervals (up to 3–4 weeks postcolonoscopy), and complication rates need to be tracked at least until this time point.

A summary of the findings of prior large studies of colonoscopy complications is shown in Table 1 . In general, the risk of colonic perforation has been most thoroughly studied. Although this is potentially the most serious complication, it is also uncommon, with reported rates from large studies of less than 0.3% and generally less than 0.1%. Lower gastrointestinal bleeding is the most common serious complication, with reported risk between 0.1% and 0.6%. Postpolypectomy syndrome, where use of electrocautery results in a transmural burn and localized peritonitis, is infrequently studied but also seems uncommon (<0.2%). Diverticulitis has also been reported as a serious complication of colonoscopy, with an incidence ranging from 0.04% to 0.08%. The reported risk of death after colonoscopy varies between 0 and 0.09%.

It is unclear whether or not the risk of complications is less with colonoscopy for screening or surveillance indications than with colonoscopy for diagnostic indications. In a systematic review, the US Preventive Services Task Force estimated the incidence of serious complications after screening colonoscopy in predominantly asymptomatic persons to be 2.8 per 1000 examinations (95% CI, 1.5–5.2). In a study of Medicare beneficiaries, Warren and colleagues found a lower risk of serious gastrointestinal complications after screening colonoscopy (2.4/1000 examinations) than after diagnostic colonoscopy (4.2/1000 examinations) or after colonoscopy with polypectomy (9.3/1000 examinations).

Less severe adverse events

Because most previous studies have focused on major complications after colonoscopy, the risks of less serious complications, such as self-limited bleeding or abdominal pain, are even less clearly documented. A substantial proportion of patients (up 33%), however, report some gastrointestinal symptoms after colonoscopy. Reported symptoms include abdominal pain (10.5%), bloating (25%), self-limited gastrointestinal bleeding (3.8%), diarrhea (6.3%), and nausea (4.0%). These symptoms generally are mild and resolve within 2 days after colonoscopy.

Less severe adverse events

Because most previous studies have focused on major complications after colonoscopy, the risks of less serious complications, such as self-limited bleeding or abdominal pain, are even less clearly documented. A substantial proportion of patients (up 33%), however, report some gastrointestinal symptoms after colonoscopy. Reported symptoms include abdominal pain (10.5%), bloating (25%), self-limited gastrointestinal bleeding (3.8%), diarrhea (6.3%), and nausea (4.0%). These symptoms generally are mild and resolve within 2 days after colonoscopy.

Sedation-related and cardiovascular complications

Serious complications from moderate sedation for colonoscopy are also uncommon but include respiratory depression, hypoxia, chest pain, cardiac arrhythmias, hypo- or hypertension, and vasovagal reactions. McQuaid and Laine, in a meta-analysis of randomized trials for moderate sedation, found a risk of hypoxemia of 18% for patients receiving midazolam alone versus 11% for patients receiving midazolam with a narcotic. In analysis of data from 174,255 colonoscopies in the Clinical Outcomes Research Initiative (CORI) database, Sharma and colleagues found an overall risk of cardiopulmonary complications after colonoscopy of 1100/100,000 procedures. The most common cardiopulmonary complications were transient hypoxia (230/100,000 procedures), bradycardia (280/100,000 procedures), hypotension (480/100,000 procedures), and vasovagal reactions (190/100,000 procedures). Oxygen supplementation was required in 7650/100,000 procedures) and intravenous fluids in 310/100,000 procedures. Reversal of sedation occurred in 490/100,000 procedures. Higher American Society of Anesthesiologists class and increasing age were significant independent predictors of cardiovascular events related to sedation, and adverse events were more common for inpatient than for outpatient procedures. The dose of midazolam used was inversely associated with risk for cardiovascular events.

Additional concerns have arisen over the potential risk of excess cardiovascular events within 30 days of colonoscopy. Warren and colleagues examined the risk of cardiovascular events in Medicare beneficiaries after colonoscopy and found a modestly increased risk of cardiovascular events requiring an emergency visit or hospitalization after colonoscopy compared with a cohort matched on age, gender, and comorbidity (unadjusted overall risk 19.4/1000 vs 16.6/1000, P <.001). The most common adverse cardiovascular event was arrhythmia (10.2/1000). This increased risk of cardiovascular events seemed most pronounced for patients who had a polypectomy (adjusted risk 23.3/1000 persons) compared with matched patients who did not undergo colonoscopy (adjusted risk 15.9/1000) or patients who underwent screening colonoscopy (adjusted risk 12.5/1000) or diagnostic colonoscopy (adjusted risk 15.8/1000). The presence of comorbidities, including diabetes, stroke, atrial fibrillation, or congestive heart failure, was associated with increased risk of cardiovascular events compared with patients without these comorbidities. The risk of cardiovascular events, however, was not significantly different from matched control groups who did not undergo colonoscopy. Another study of screening and surveillance colonoscopy, with generally younger patients, did not find an excess risk of cardiovascular events.

Complications related to bowel preparation

Bowel preparation regimens fall into two general categories: electrolyte solutions containing polyethylene glycol (PEG) (eg, GoLYTELY, Colyte, and NuLYTELY) and non-PEG solutions, including sodium phosphate solutions (eg, Fleet’s Phospho-Soda, OsmoPrep, and Visicol). Although lower-volume sodium phosphate preparations are often better tolerated, concerns have arisen about renal injury after these preparations. These preparations are hyperosmotic and can lead to fluid retention, leading to concerns about their use in patients with underlying cardiac, renal, or hepatic diseases. Studies of healthy adults undergoing colonoscopy have shown development of significant hypocalcemia or hyperphosphatemia. Most concerning is the risk of developing acute phosphate nephropathy due to precipitation of calcium-phosphate crystals in the kidney. The risk of acute phosphate nephropathy may be increased in the elderly or in those who are inadequately hydrated. Risk may also increase with hypertension or with use of certain antihypertensive medications, including diuretics, angiotensin-converting enzyme (ACE) inhibitors, or angiotensin receptor blockers (ARBs).

The overall risk of acute phosphate nephropathy is unknown but is estimated at less than 0.1%. In a retrospective study, Hurst and colleagues found acute kidney injury, defined as an increase in serum creatinine of greater than 0.5 mg/dL, in 1.2% of patients within 12 months of colonoscopy. Acute kidney injury was more common in patients who had received a sodium phosphate preparation compared with PEG-based solutions (1.3% vs 0.9%, P <0.001). Additional risk factors for acute kidney injury in this study included older age and congestive heart failure. Brunelli and colleagues found a risk of acute kidney injury only in patients who are also receiving ACE inhibitors or ARBs. Despite these concerns about sodium phosphate preparations, other retrospective studies have found similar proportions of patients who develop renal insufficiency after colonoscopy with sodium phosphate or PEG-based preparations.

Because of these reports, oral sodium phosphate solutions have fallen out of favor for bowel preparation, and the Food and Drug Administration has issued a safety warning about these drugs. Use of sodium phosphate preparations is not advised in elderly patients, those with pre-existing kidney disease, and those with pre-existing fluid or electrolyte imbalances, such as patients with congestive heart failure. The importance of adequate hydration with these preparations has also been emphasized.

PEG-based solutions do not lead to substantial shifts in fluid levels and can be safely used for patients with electrolyte imbalances, congestive heart failure, or chronic kidney disease. Because of the large volume of the preparation required, however, they are often difficult for patients to tolerate. Symptoms of nausea/vomiting and abdominal fullness are common during ingestion. Less common complications associated with large-volume preparations include vomiting-induced Mallory-Weiss tears, esophageal rupture, pulmonary aspiration, hypothermia, and cardiac arrhythmias. Although reported, electrolyte abnormalities are less common with PEG-based solutions than with sodium phosphate preparations.

Miscellaneous complications

Miscellaneous and rare complications of colonoscopy include splenic hematoma or rupture, acute appendicitis, incarcerated hernias, subcutaneous emphysema in the absence of perforation, intramural hematoma, and ischemic colitis. Colonic explosions have been reported with electrocautery in patients with suboptimal bowel preparation, in particular with mannitol preparations, which are not currently used. Although transient bacteremia occurs frequently after colonoscopy, complications of bacteremia, such as infective endocarditis, are rare, and antibiotic prophylaxis is not routinely recommended. If glutaraldehyde is used in the endoscope disinfection process, a chemical colitis can occur if the endoscope has not been adequately rinsed before the next procedure.

Risk factors for serious colonoscopy complications

Increasing age has consistently been found a significant risk factor for serious gastrointestinal events after colonoscopy. For example, Gatto and colleagues found a nearly 4-fold increased risk of perforation for subjects aged 75 to 79 years compared with those aged 65 to 69 years (odds ratio [OR] 3.5; 95% CI, 1.5–7.8). Female gender may be a risk factor for colonic perforation in particular. This may related to the finding that colonoscopy is generally more difficult in women than in men, likely due to pelvic anatomy and prior pelvic surgery, which is more common in women. Although the presence of comorbid conditions, such as diabetes, stroke, atrial fibrillation, and congestive heart failure, has been associated with increased risk of serious adverse events in the Medicare population, this association has not been found consistently in all studies.

The risk of complications increases markedly in examinations where a polypectomy is performed, in particular with use of electrocautery. Levin and colleagues found a 9-fold increased risk of any serious complication in colonoscopies where a biopsy or polypectomy was performed (OR 9.2; 95% CI, 2.9–29.0), although the risk of perforation alone was not significantly increased (OR 1.9; 95% CI, 0.3–10.9). The risk of complications, most commonly gastrointestinal bleeding, increases further if more than one polypectomy with electrocautery is performed and with larger polyp size. Biopsy alone may also increase the risk of complications. Prior studies have suggested that hot biopsy for removal of polyps is associated with increased bleeding risk, particularly if used in the proximal colon.

Although some investigators hypothesize that endoscopist experience and skill may influence complication rates, there are few data to support or refute this possibility. In Canada, endoscopist specialty did not influence the risk of complications, but endoscopists with low colonoscopy volumes had higher risks of complications than high volume providers (OR 2.96 comparing lowest to highest quintile; 95% CI, 1.57–5.61). When only including colonoscopies performed by gastroenterologists, however, colonoscopy volume was not associated with risk of complications. Studies are needed in other settings and countries to confirm these findings.

The use of warfarin or clopidogrel may also increase risk of colonoscopy complications, in particular postpolypectomy bleeding. For example, Hui and colleagues found that warfarin increased the risk of postpolypectomy bleeding (OR 13.37; 95% CI, 4.10–43.65), even after adjusting for other potential risk factors, such as patient age, location and size of polyp, polypectomy technique, and presence of underlying renal impairment. In contrast, use of aspirin or nonsteroidal anti-inflammatory medications has not been shown to increase the risk of postpolypectomy bleeding. There are few published data about the risk of bleeding complications after use of clopidogrel.

Published guidelines recommend that management of anticoagulants and antiplatelet agents be individualized according to procedure risk and risks associated with the underlying condition. For example, patients on warfarin therapy with low-risk conditions (such as nonvalvular atrial fibrillation, bioprosthetic valves, and mechanical valves in the aortic position) undergoing high-risk procedures (such as colonoscopy with polypectomy) should discontinue warfarin 3 to 5 days before the procedure. In patients with high-risk conditions (such as mechanical valves in the mitral position or valvular atrial fibrillation) undergoing high-risk procedures, warfarin should be discontinued 3 to 5 days before the procedure with consideration of bridging therapy with low molecular weight heparins. The effect of the timing of reinstitution of warfarin therapy on complication risk is also not well studied, but resumption of anticoagulation within 1 week of polypectomy seems to increase the risk of gastrointestinal bleeding (OR 5.2; 95% CI, 2.2–12.5). Current guidelines recommend reinstituting therapy on the same night as the procedure, however, although delays can be considered if the polyps removed are large or otherwise at high risk for bleeding.

American Society of Gastrointestinal Endoscopy guidelines also suggest that aspirin or nonsteroidal anti-inflammatory drugs do not need to be discontinued for patients undergoing endoscopic procedures. Decisions about discontinuation of clopidogrel also need to be individualized according to the underlying indications for its use and the risks of the planned procedure.