Bowel preparation agent inducing profound shock precolonoscopy

Abstract

A 73-year-old woman was admitted to the intensive care unit following vomiting and diarrhoea onset after completing oral bowel preparation prior to colonoscopy to investigate haematochezia. She had a history of severe chronic obstructive pulmonary disease, Crohn’s disease, diverticular disease, hypertension and dyslipidaemia. She was resuscitated with intravenous fluids, antibiotics and required epinephrine, norepinephrine and vasopressin infusions. She improved over her 4-day intensive care admission and was discharged to the general medical ward, but ultimately died 19 days after presentation.

Background

Colonoscopy is a commonly performed procedure accepted as carrying relatively low risk, with a reported mortality of 0.007%.¹ Often overlooked however are the risks associated with bowel preparation, with fluid and electrolyte shifts potentially causing significant derangement to physiology. The elderly, patients with heart failure, kidney disease, liver disease, pregnancy, inflammatory bowel disease or pre-existing electrolyte abnormalities are particularly at risk.

This case presents us with a patient who became profoundly shocked post completion of bowel preparation regimen. Vomiting, diarrhoea and fluid shifts likely explain the hypovolaemic component of shock, however the patient’s condition may also have been explained by an element of septic shock. Transient bacteraemia has been reported during and after colonoscopy due to bacterial translocation of normal colonic flora to the blood stream² and there are also reports of sepsis precipitated by the preparation itself.

Ultimately this case highlights the risks associated with bowel preparation prior to colonoscopy and seeks to remind clinicians of the potential complications to be wary of when consenting potential patients.

Case presentation

The patient presented to the emergency department with severe nausea and vomiting onset less than 1 hour after finishing sodium picosulfate/magnesium oxide/citric acid bowel preparation prior to colonoscopy. She was well prior to commencing the regimen. Relevant history included severe chronic obstructive pulmonary disease, Crohn’s disease, diverticular disease, cardiomyopathy, hypertension and dyslipidaemia. She was immunosuppressed on azathioprine for her inflammatory bowel disease. Other medications included aspirin, bisoprolol, perindopril, pantoprazole, salbutamol, fluticasone/salmeterol and rosuvastatin. She had no surgical history apart from prior colonoscopies. She was usually dependent on her husband for fulltime cares and only able to mobilise short distances with a four-wheel walker. Her American Society of Anesthiologists classification was a 3.

On arrival she was shocked with an unrecordable blood pressure and oxygen saturations, temperature of 33° and heart rate of 65. She was cool peripherally and confused in her speech. Her abdomen was firm and generally tender. Intravenous access was gained and she was resuscitated with 3 L of intravenous balanced crystalloid, intravenous antiemetics, piperacillin/tazobactam and gentamicin. Peripheral norepinephrine and epinephrine infusions were initiated to maintain blood pressure and counter persistent bradycardia. CT scan showed non-specific stranding along the descending colon but no acute surgical pathology. She was transferred to the intensive care unit (ICU) for ongoing management.

Investigations

Initial venous blood gas in the emergency department showed a pH of 7.36, alveolar carbon dioxide tension of 60 mm Hg, bicarbonate of 34 mmol/L, base excess of 6.7 mmol/L and lactate of 4.9 mmol/L. Full blood count displayed a haemoglobin of 136 g/L, platelets of 310×10⁹/ L, white cell count of 15.5×10⁹/ L and neutrophils of 13.75×10⁹/ L. Sodium was 132 mmol/L, potassium 4.1 mmol/L and magnesium was elevated at 2.84 mmol/L. Renal function studies showed a urea of 5.1 mmol/L, creatinine of 106 umol/L and estimated glomerular filtration rate of 45 mL/min/1.73 m². Lipase was within normal limits at 30 U/L and urine sample was not consistent with urinary tract infection. All blood cultures returned negative.

Differential diagnosis

Differential diagnosis on presentation was initially wide. The most obvious cause of her shock was hypovolaemia secondary to gastrointestinal losses precipitated by bowel preparation. Her partial responsiveness to fluid resuscitation as well as the profound nature of her shock, requiring multiple vasoactive agents suggests there were likely concurrent mechanisms contributing to her presentation. There was a cardiogenic component given she was persistently bradycardic during resuscitation, however focused cardiac ultrasound displayed adequate cardiac filling, preserved left and right ventricular systolic function, no pericardial effusion, no major valvular pathology and no regional wall motion abnormalities. Similarly, there did not appear to be an obstructive cause for her condition. Given her significant vasoplegia, widened pulse pressure and neutrophilia, in addition to aforementioned findings, there was postulated to be a septic component, presumed secondary to bacterial translocation sustained during the preparation regimen. She also appeared to improve significantly following the initiation of antimicrobials. However, blood cultures and stool samples returned negative and so causative organisms were never identified. Ischaemic colitis secondary to hypotension on a background of chronic hypoxia was another consideration. CT scan did not display any overt intra-abdominal pathology, such as haemorrhage or perforation, to help explain her condition.

Treatment

Given the patient’s comorbidities an advanced resuscitation plan was discussed with the patient and family members and a ceiling of care established. It was agreed to treat with intravenous fluids and antibiotics, vasoactive agents and non-invasive ventilation if required, however the patient would not be for intubation and ventilation, renal replacement therapy, CPR or defibrillation. Supportive treatment was undertaken in the intensive care unit. Oxygen supplementation in the form of high flow nasal prongs was used to assist with work of breathing and improve lung recruitment. Intravenous fluid resuscitation was continued with 4% albumin. An epinephrine infusion was continued to optimise heart rate and cardiac output, while norepinephrine and vasopressin infusions were utilised to sustain a mean arterial pressure of greater than 65 mm Hg. Antimicrobials consisted of piperacillin/tazobactam, metronidazole and gentamicin. Six hourly doses of stress hydrocortisone were initiated. In addition to the above she received regular ICU supportive cares, including intravenous pantoprazole, prophylactic heparin and tight blood sugar control. She was reviewed by the general surgical team who concluded that there was no surgical intervention indicated.

Outcome and follow-up

The patient initially improved during her ICU admission. Epinephrine, norepinephrine and vasopressin infusions were weaned and she was discharged to the medical ward after a 4-day admission. She was stable on the medical ward for 6 days before developing a small bowel obstruction. She was deemed to not be a surgical candidate and thus was managed conservatively with nasogastric tube decompression. She continued to deteriorate on the ward after which it was agreed to switch to comfort cares. The patient ultimately passed away 19 days after presentation.

Discussion

In order to provide adequate informed consent for a procedure clinicians are required to discuss relevant and important associated risks with the patient. When consenting patients for colonoscopy, the relevant risks including perforation, bleeding and incomplete studies are often discussed. However, the risks of bowel preparation itself are commonly overlooked. The purpose of this case report is to highlight the risks of bowel preparation, which while usually less common and serious than the endoscopy, can potentially have disastrous consequences. This report in particular describes a patient who sustained severe shock following bowel preparation. Oral bowel preparation solutions have a number of recognised risks linked to fluid and electrolyte shifts including increases in plasma volume exacerbating heart failure^{3 4} and hyponatraemia.⁵ Rarer side effects include aspiration, Mallory-Weiss tears, oesophageal perforation, pancreatitis, cardiac dysrhythmia and exacerbation of syndrome of inappropriate antidiuretic hormone secretion.^6–9

Few case reports however describe patients who become profoundly shocked following bowel preparation. One report describes a novel case in which a patient presented with lethal hypermagnesaemia precipitating severe cardiovascular collapse, neurological depression and acute nephropathy following completion of preparation containing magnesium citrate.¹⁰ Our patient’s magnesium level peaked at 3.12 mmol/L, significantly above the normal reference range, however much lower than previously described lethal levels (>7 mmol/L).

While unproven objectively, it was postulated that our patient likely had an element of septic shock complicating her presentation. There are rare cases of patients developing sepsis following bowel preparation. One case described a patient who developed septic shock immediately following completion of the preparation requiring aggressive resuscitation.¹¹ This patient had proven bacteraemia in the form of blood cultures positive for Citrobacter braakii. Like our patient the indication for colonoscopy was to investigate haematochezia and authors also hypothesised that sepsis occurred to due bacterial translocation, however it was difficult to definitively prove. Another reported a patient with a history of ulcerative colitis who developed sepsis 2 hours after completing bowel preparation, with cultures positive for Escherichia coli.¹² Again, this draws similarities to our patient with Crohn’s disease. While rare it appears that bowel preparation has the potential to precipitate bacteraemia and sepsis in particular patients.

The final major learning point from this case centres on the importance of weighing up the risks of a procedure against the potential benefits, and whether the patient would be able to tolerate a complication and any intervention if this was to occur. Given the accepted low risk of bowel preparation and the rarity of becoming critically unwell it is unreasonable to suggest that this patient not be offered a colonoscopy on this basis alone. However, if a more common side effect, such a perforation during the procedure had occurred, her comorbidities would likely have limited the available surgical intervention or made a procedure extremely high risk. Given this risk it could be argued that she did not require the investigation at all. Alternatively, a less invasive method, such as CT colonography with faecal tagging could have been employed. I encourage the reader to take into account a patient’s comorbidity burden and physiological reserve when assessing their suitability for a procedure and consider whether they would be able to tolerate a potential complication.

Patient’s perspective.

We just really think that mum’s doctor should have taken into consideration her health issues before booking her in for a colonoscopy. Of course it was just bad luck that she got a bowel blockage after trying to recover from the septic shock. We miss our mum dearly and I am hoping that this report can help people in the future.

Learning points.

• While less common than complications due to the endoscopy itself, consider the potential complications of bowel preparation when assessing a patient’s suitability for colonoscopy.

• Patients most at risk of complications from bowel preparation include those with heart failure, renal insufficiency, chronic liver disease, electrolyte abnormalities, pregnancy and inflammatory bowel disease.

• Potential complications include fluid and electrolyte shifts, heart failure exacerbation, pancreatitis, oesophageal perforation, cardiac dysrhythmia and exacerbation of syndrome of inappropriate antidiuretic hormone secretion.

• In rare cases bowel preparation may precipitate sepsis from bacterial translocation from the gastrointestinal tract.

• All patients are different, consider risks vs benefits of a procedure on a particular patient and whether they would be able to tolerate a potential complication.