Postoperative Day Five: A Silent Abdomen in a Steroid-Treated Diabetic
A 62-year-old woman is on day 5 after an elective sigmoid colectomy with a stapled colorectal anastomosis. She has type 2 diabetes (HbA1c 8.9%) and takes prednisolone 10 mg daily for polymyalgia rheumatica. The nurses report that she is "not herself". Temperature 38.4 °C, heart rate 124/min, blood pressure 96/58 mmHg, respiratory rate 26/min, SpO2 94% on air, and urine output 20 mL/hour for the last 3 hours. The abdomen is distended with diffuse tenderness but no rigidity. White cell count is 3.1 x 10^9/L, lactate 3.6 mmol/L, creatinine 168 micromol/L (baseline 74). She has received intravenous cefuroxime and metronidazole since induction. Chest radiograph shows no consolidation and urine dipstick is negative.
1. What is your working diagnosis and what will you do in the first hour? Explain the significance of her white cell count and steroid use.
Read the worked answer
Working diagnosis. Sepsis with a probable intra-abdominal source, most likely an anastomotic leak, in an immunocompromised, poorly controlled diabetic host. The negative chest film and urine dipstick make pneumonia and urinary sepsis unlikely, and day 5 is the classic window for a leak.
First hour:
- Recognise and escalate. This meets sepsis criteria with hypoperfusion (lactate 3.6, oliguria, rising creatinine). Call for senior surgical review and anaesthetic support immediately.
- Cultures then antibiotics, without delay. Two sets of blood cultures and urine culture, but do not let a difficult venepuncture delay antimicrobials beyond about 45 minutes. She is already on cefuroxime and metronidazole - that regimen will not cover resistant Gram-negatives or enterococci in a patient who has had five days of antibiotics.
- Resuscitation. Balanced crystalloid 30 mL/kg in aliquots, guided by response, with early vasopressors if she remains hypotensive after the initial bolus; urinary catheter for hourly output; arterial line.
- Send lactate, full blood count, coagulation, glucose and a venous blood gas and correct hyperglycaemia with an insulin infusion.
The leucopenia of 3.1 is a warning sign, not reassurance: it suggests a consumptive, overwhelming process and is seen with severe sepsis, particularly in patients on corticosteroids. Prednisolone blunts fever, tachycardia and peritoneal signs - the absence of rigidity is meaningless here, and her relative bradycardia-free presentation is a falsely reassuring picture. I would treat the clinical trajectory and the lactate, not the abdominal examination.
2. Which imaging would you request, and how would you decide between imaging and immediate laparotomy?
Read the worked answer
Imaging of choice: CT of the abdomen and pelvis with intravenous contrast, provided she can be stabilised for the 10–15 minutes it takes. It identifies free fluid, free gas, anastomotic dehiscence, abscess, ischaemia and collection site, and it guides whether percutaneous drainage is possible. Oral contrast is unnecessary and unhelpful in this setting; a water-soluble contrast enema is reserved for a patient who is stable with equivocal CT and a suspected low anastomotic leak.
Decision rule. CT is for the patient who responds to initial resuscitation and in whom the answer will change management - drainage versus laparotomy versus continued antibiotics. It is not for the patient who remains hypotensive with a rising lactate and peritonism; that patient goes straight to theatre without leaving the resus room.
What I would fear most. Anastomotic dehiscence with faecal peritonitis. If CT shows free gas, contrast extravasation or a large collection with ongoing sepsis, the treatment is source control: re-laparotomy, take down the anastomosis, washout and end colostomy. Antibiotics alone do not fix a hole.
I would also keep a low threshold for theatre if her lactate fails to clear by 2–4 hours despite fluids, since delay beyond about 6 hours in a leak markedly increases mortality.
3. She undergoes laparotomy for a leaking anastomosis. Outline your antimicrobial plan for the next 48 hours and your approach to stopping or changing antibiotics, stating the reasoning behind each choice.
Read the worked answer
Empirical escalation is justified. At laparotomy there is faecal contamination and established sepsis; start piperacillin-tazobactam (or, if she has had recent extended-spectrum cephalosporins and healthcare exposure, consider a carbapenem) plus metronidazole if not already covered. The previous cefuroxime/metronidazole combination is inadequate for this level of contamination.
Consider antifungal cover. Risk factors here are multiple: prolonged antibiotics, diabetes, steroid use, major gastrointestinal surgery and the need for prolonged critical care. If she has a high candida colonisation index, needs total parenteral nutrition, or has recurrent leaks, add an echinocandin or fluconazole.
Then narrow deliberately. Take deep wound and peritoneal cultures at operation. At 48–72 hours:
- If cultures grow a susceptible organism and she is improving - de-escalate to the narrowest agent that covers it; there is no evidence that longer broad-spectrum cover improves outcome in source-controlled peritonitis.
- If nothing grows and the source is controlled - stop the antifungal agent.
- If she is not improving - re-image for undrained collections rather than simply broadening the spectrum. Escalating antibiotics without addressing a residual collection is the most common error.
Duration should be governed by adequacy of source control, typically 4–7 days of therapy after adequate drainage, with clinical or biochemical (procalcitonin, white cell count, lactate) guidance for stopping. Her diabetes must be actively controlled; hyperglycaemia perpetuates infection.