Inpatient Glycaemic Control After the 2017 Protocol Debate

The debate over inpatient glycaemic control in May 2017 was less about finding a single “best” insulin regimen and more about deciding which patients needed intensive treatment, how quickly glucose should be corrected, and how safely hospitals could deliver a consistent plan. At Hospital Medicine 2017 in Las Vegas, hospitalists, pharmacists, nurses and diabetes specialists examined the practical gap between guideline recommendations and the unpredictable reality of acute care.

That discussion remains relevant for Australian hospitals. A patient admitted to a metropolitan emergency department in Melbourne, a regional hospital in Townsville or a private facility in Perth may arrive with different medicines, dietary patterns, renal function and access to endocrinology support. The central challenge is the same: control hyperglycaemia without creating avoidable hypoglycaemia, treatment delays or confusion at the point of discharge.

Why insulin protocols were under scrutiny

Traditional sliding-scale insulin was increasingly criticised because it reacted to high capillary glucose rather than anticipating the patient’s basal insulin requirement. A correction dose given after a reading of 15 mmol/L might lower the number temporarily, yet leave the person exposed to recurring hyperglycaemia between meals and overnight. For patients with type 1 diabetes, omitting basal insulin could create a serious risk of ketosis even when they were eating very little.

The preferred alternative in many discussions was a scheduled basal-bolus regimen. Long-acting insulin supplied background coverage, rapid-acting insulin was matched to meals, and correction insulin addressed additional elevations. This approach could provide smoother control, but it required reliable meal timing, regular bedside glucose testing and staff who understood how to adjust doses when appetite, steroids or kidney function changed.

In Australia, hospital policies commonly record blood glucose in mmol/L rather than the mg/dL units used in many American publications. That difference is clinically important when protocols are adapted from overseas. A target expressed as 140–180 mg/dL is approximately 7.8–10.0 mmol/L, so local order sets must use clear units and avoid ambiguous handwritten conversions.

Choosing the right target for acute illness

The 2017 conversation reflected a move away from intensive glucose lowering for every critically ill patient. Earlier enthusiasm for near-normal glycaemia had been tempered by evidence that aggressive intravenous insulin could produce severe hypoglycaemia without improving outcomes in a broad intensive care population. For many non-critical inpatients, a moderate target was considered safer than pursuing normal glucose at any cost.

Targets still needed to be individualised. A young adult with type 1 diabetes, a person receiving high-dose corticosteroids for an asthma exacerbation, and an older patient with dementia and chronic kidney disease should not automatically receive the same correction scale. Frailty, nutritional intake, infection severity, renal clearance and previous hypoglycaemic events all influence the balance between benefit and harm.

The language used in a protocol can shape clinical behaviour. “Tight control” may encourage staff to chase every elevated result, while a target range with explicit escalation rules gives clinicians room to assess the whole patient. A useful order set identifies when to call the medical team, when to reduce insulin after a poor meal, and how to respond to glucose below the hospital’s defined safety threshold.

Basal-bolus, correction insulin and intravenous therapy

For a stable medical ward patient who is eating, basal-bolus therapy generally offered a more physiological approach than correction-only insulin. The initial total daily dose had to be conservative, particularly for insulin-naive patients, older adults and those with reduced renal function. Dividing the dose between basal and prandial components allowed clinicians to pause meal insulin when food was withheld while maintaining essential background coverage.

Correction doses remained useful, but they worked best as an addition rather than the entire treatment plan. A scale based on insulin sensitivity was preferable to a universal dose. A patient already receiving substantial corticosteroids might need more insulin during the day, whereas a patient with declining renal function could require rapid dose reduction. Reassessment after several readings was more reliable than allowing an unchanged charted scale to continue for days.

Intravenous insulin was reserved for situations requiring close, rapid control, such as diabetic ketoacidosis, hyperosmolar hyperglycaemic state and selected critically ill patients. It demanded a dedicated infusion protocol, frequent glucose checks, potassium surveillance and a clear transition plan. Before stopping the infusion, subcutaneous basal insulin usually needed to overlap for an appropriate period so that insulin deficiency did not recur.

The conference archive provides useful historical context for the education and clinical exchange surrounding these decisions. The value of such material lies in seeing insulin safety as a systems issue involving prescribing, nursing observation, pharmacy review and discharge communication rather than as a dose calculation alone.

Preventing hypoglycaemia across the ward

Hypoglycaemia was a major reason clinicians questioned aggressive inpatient protocols. A dose that appears reasonable at breakfast may become excessive if a scan delays lunch, a patient vomits, renal function deteriorates or a steroid is discontinued. Hospitals therefore needed procedures that connected insulin administration with meal delivery, fasting status and changes in clinical condition.

Australian wards face familiar operational pressures: a patient may wait in an emergency department cubicle, transfer between units at Royal Brisbane and Women’s Hospital or a Sydney teaching hospital, and then miss a meal during imaging. Protocols should state who is responsible for checking whether food is available before rapid-acting insulin is given. They should also make the treatment of low glucose immediately visible, with standard oral or intravenous options and a documented recheck.

Medication reconciliation was equally important. Patients may use multiple daily injections, premixed insulin, insulin pumps or newer glucose-lowering medicines before admission. SGLT2 inhibitors had already raised concern in relation to euglycaemic ketoacidosis, particularly during fasting and acute illness. A safe admission process records the last dose, home device settings, usual glucose pattern and whether the patient can manage injections independently.

Pharmacist involvement can make these safeguards more dependable. A pharmacist reviewing insulin orders alongside a hospitalist can identify duplicate correction scales, inappropriate continuation of oral medicines and discrepancies between the medication chart and the discharge plan. The practical model described in this collaborative rounding guide is especially relevant when several teams share responsibility for a complex inpatient.

Making protocols work at discharge

An inpatient insulin protocol should begin with a discharge destination in mind. A temporary basal-bolus regimen may be suitable during pneumonia treatment but excessive once corticosteroids stop and normal meals resume. Conversely, a person newly diagnosed with diabetes may leave hospital needing education, equipment, prescriptions and a prompt review appointment rather than a vague instruction to “see your doctor”.

Australia’s public system adds practical considerations. Patients may obtain diabetes supplies through the National Diabetes Services Scheme, while availability and familiarity can vary in remote communities. A person returning to a rural area outside Darwin or travelling several hours from Adelaide may not have immediate access to an endocrinologist or diabetes educator. Discharge instructions should therefore include sick-day guidance, hypoglycaemia treatment, insulin storage, injection technique and a named follow-up pathway.

Documentation should use local units, plain language and an explicit medication schedule. It should state whether insulin is to be taken with food, what to do if a meal is missed, and when to seek urgent care. A written plan shared with the general practitioner, community nurse, residential aged-care facility or Aboriginal Community Controlled Health Service can prevent the common failure in which the inpatient chart is detailed but the next clinician receives almost no usable information.

The most durable lesson from the 2017 debate was that a protocol is a safety framework, not a substitute for judgement. Hospitals need standard starting doses and escalation pathways, but they also need permission to adapt treatment when a patient’s intake, renal function, infection or medication exposure changes. Auditing severe hypoglycaemia, persistent hyperglycaemia, missed insulin doses and readmissions can show whether the protocol is improving care in practice.

A sensible ward approach is therefore straightforward: continue essential basal insulin, use scheduled therapy when appropriate, reserve intravenous insulin for clearly defined indications, match doses to nutrition and illness, and build pharmacy and nursing checks into every transition. In everyday practice, the safest glycaemic protocol is the one that makes the right action clear at 2 a.m., during a delayed meal, and on the morning the patient goes home.