Tachycardia and Rising Creatinine on Day Four After Sigmoid Colectomy
A 64-year-old man with type 2 diabetes (HbA1c 8.4%) and a BMI of 33 underwent laparoscopic sigmoid colectomy with a stapled primary anastomosis four days ago for perforated diverticulitis. Overnight the nurse documents temperature 38.1 °C, heart rate 118/min, BP 104/62 mmHg, respiratory rate 22/min, and urine output of 20 mL/h for the last four hours. He has received 2.5 L of crystalloid since surgery. The abdomen is mildly distended and tender in the left iliac fossa without guarding or rigidity; the drain has put out 150 mL of serous fluid today. Bloods: creatinine 168 micromol/L (baseline 88), lactate 2.4 mmol/L, white cells 15.8, platelets 210. The chest is clear, there is no calf tenderness, and he is on ceftriaxone and metronidazole. The surgical registrar is off site and the CT scanner is available.
1. Give your leading diagnosis and the two alternative explanations you must exclude. State the investigation that discriminates best and explain why a bedside fluid challenge alone is unreliable here.
Read the worked answer
Leading diagnosis: established intra-abdominal sepsis, most probably an anastomotic leak. The combination of fever, tachycardia, four hours of oliguria, creatinine nearly doubled and a lactate of 2.4 in an obese diabetic four days after a stapled colorectal anastomosis sits squarely in the classic window for a leak.
Alternatives to exclude. (a) Hypovolaemia or third-space loss - 2.5 L over four days is modest, and a serous drain does not prove otherwise. (b) Non-abdominal sepsis - atelectasis, line sepsis, urinary infection or Clostridioides difficile - far more common and far more benign on day four.
Best discriminator. CT abdomen and pelvis with intravenous contrast plus oral contrast (and rectal contrast if the anastomosis is low): it shows free fluid, extraluminal gas, the anastomotic defect and any collection, and it can guide percutaneous drainage. Blood and urine cultures, lactate trend and a bedside ultrasound are adjuncts, not substitutes.
Why a fluid challenge alone fails. In a leak, a 250-500 mL bolus can transiently normalise the pulse and urine output by improving preload while the septic source persists, giving false reassurance and delaying source control; conversely, a non-responder may simply be hypovolaemic. If you do challenge, make it a defined 250-500 mL of balanced crystalloid with reassessment at 30 minutes, and never let it replace imaging.
2. Set out your management for the next six hours, and specify the findings that would take him to theatre without waiting for imaging.
Read the worked answer
First hour. Take blood cultures before antibiotics, then give broad-spectrum cover - piperacillin-tazobactam or, in this unstable diabetic, a carbapenem plus Gram-positive cover per local policy; ensure metronidazole anaerobic cover is retained. Insert a urinary catheter to measure output, start a 500 mL balanced crystalloid bolus with reassessment, send lactate, FBC, coagulation and cross-match, and correct hyperglycaemia with an insulin infusion.
Imaging. CT within the hour, with the images reviewed by a surgeon rather than relying on the report alone. Involve the consultant now, not after the scan.
Theatre without waiting for CT if: he develops generalised peritonitis or rigidity, has free intraperitoneal air, becomes haemodynamically unstable after 30 mL/kg of crystalloid, has a rising lactate, or if a drain starts to pass faecal material. In an unstable patient with a suspected free leak, washout and defunctioning is the priority; percutaneous drainage is appropriate only for a contained, stable collection.
If a leak is confirmed: take down the anastomosis, perform thorough lavage and either a Hartmann procedure or, with healthy tissue and limited contamination, repair with a defunctioning stoma - the choice turns on contamination, tissue quality and stability. Book ICU post-operatively and inform the family.
3. A colleague argues that a lactate of 2.4 mmol/L and the absence of rigidity exclude a leak, and proposes waiting for a routine CT in the morning. Respond to that reasoning.
Read the worked answer
Argue from the trajectory, not a single value. A lactate of 2.4 mmol/L is abnormal, and it sits alongside a creatinine that has almost doubled, four hours of oliguria and sustained tachycardia. The composite picture is the signal; each number alone is dismissible.
Rigidity is a late, insensitive sign. Early anastomotic leak produces pelvic, not generalised, peritoneal irritation, and in an obese, diabetic abdomen after laparoscopy, guarding is unreliable. Waiting for rigidity is waiting for a catastrophe.
The drain is not reassuring. A small proximal leak can produce clear serous effluent; the volume and character of drain output have poor sensitivity and specificity.
Confounders cut both ways. Diabetes and any metformin, plus hypovolaemia and tissue hypoperfusion, can raise lactate without a leak - which is exactly why lactate cannot be used to exclude one either.
Decision rule. In a post-colectomy patient who is not improving after adequate resuscitation, a leak remains the working diagnosis until imaging or the clinical course excludes it. The threshold for CT should be low, and if the patient is unstable the threshold for theatre should be lower still - physiology first, imaging second. A six-hour delay in source control in established sepsis measurably increases mortality.