Isophane insulin (NPH)
NPH insulin, neutral protamine Hagedorn insulin, isophane insulin human, human insulin isophane suspension, intermediate-acting human insulin
Isophane insulin, universally known as NPH, is not an insulin analogue at all. The insulin molecule in it is unmodified human insulin, identical in sequence to what the pancreas makes. What is engineered is the formulation: the insulin is crystallised with protamine, a small arginine-rich protein from fish sperm, plus zinc, producing a cloudy suspension that must be resuspended before injection and from which insulin dissolves slowly over 12 to 18 hours. Devised by Hans Christian Hagedorn in Denmark in 1936, it remains the cheapest basal insulin in the world and is available without prescription in most of the United States.
Mechanism
The active molecule is human insulin: a 51-residue two-chain polypeptide with the sequence of the endogenous hormone, produced recombinantly in Escherichia coli or Saccharomyces cerevisiae. All the pharmacological engineering is in the physical formulation rather than the chemistry. Protamine sulfate is a strongly basic, arginine-rich protein historically obtained from salmon or other fish sperm. Mixed with insulin and zinc at neutral pH in a defined stoichiometric ratio (the 'isophane' ratio, meaning neither component is in excess), the positively charged protamine and negatively charged insulin co-crystallise into rod-shaped microcrystals suspended in the vial. This is why the product is cloudy and why it must be gently resuspended before every dose; incomplete resuspension is a well-documented and entirely avoidable cause of erratic glucose control.
After injection, tissue proteases slowly degrade the protamine, releasing insulin hexamers which then dissociate to dimers and monomers and are absorbed. That dissolution step is what protracts the action: onset at one to two hours, a distinct peak at four to ten hours, and a duration of 12 to 18 hours. Unlike glargine (precipitation), detemir and degludec (fatty acid acylation and albumin binding) or icodec (albumin binding plus reduced receptor affinity), the protraction here is a matter of crystal chemistry and enzymatic release. The pronounced peak is the clinical weakness: it does not match physiological basal secretion, which is essentially flat, and when a bedtime dose peaks at three or four in the morning it is a direct cause of nocturnal hypoglycaemia. Once absorbed, the insulin is ordinary human insulin, acting at the insulin receptor with normal affinity and a normal metabolic-to-mitogenic ratio, signalling through IRS-1/2, PI3K and AKT to translocate GLUT4 into muscle and adipocyte membranes, to suppress hepatic gluconeogenesis and glycogenolysis, to inhibit lipolysis and to promote protein synthesis.
What the research shows
Isophane insulin occupies an unusual position: it has the best long-term outcome evidence of any insulin and the weakest marketing case. The DCCT and UKPDS 33 (the two trials that established that lowering glucose prevents retinopathy, nephropathy and neuropathy) both used human insulin, and no analogue has ever been tested against a comparable outcome endpoint. Every claim that analogues 'improve diabetes outcomes' is an extrapolation from HbA1c, and Cochrane reviewers have said so explicitly.
Where isophane insulin genuinely underperforms is in its pharmacokinetic shape. The four-to-ten-hour peak does not resemble physiological basal secretion, and a bedtime dose peaking in the early hours is a mechanical cause of nocturnal hypoglycaemia. Randomised trials consistently show roughly a quarter fewer nocturnal episodes with glargine or detemir at the same HbA1c, and the Riddle treat-to-target trial quantified this precisely while confirming that HbA1c itself was identical. It also requires resuspension, which introduces a real-world source of variability that trial conditions understate. The Lipska cohort of 25,489 patients complicates the analogue case considerably: in routine care, without the structured titration support of a trial, initiating an analogue was not associated with fewer hypoglycaemia-related emergency attendances, and HbA1c was marginally better with isophane insulin. That is observational evidence with the usual limitations, but it is large and it points the opposite way to the marketing. The defensible summary (and the one NICE has adopted for type 2 diabetes) is that isophane insulin is the appropriate first choice on cost-effectiveness grounds, with analogues reserved for people with recurrent or troublesome nocturnal hypoglycaemia, those needing help with injections who benefit from once-daily dosing, or those who cannot manage a suspension.
Evidence assessment
High-quality evidence
Isophane insulin was the basal insulin used in both landmark trials that established the relationship between glycaemic control and diabetes complications: the DCCT in type 1 diabetes (n=1,441, mean 6.5 years) and UKPDS 33 in type 2 diabetes (n=3,867, median 10 years). No insulin analogue has an outcome evidence base of that quality. A Cochrane review and a large real-world cohort further establish that analogues do not outperform it on HbA1c or on serious hypoglycaemia in ordinary practice, though randomised trials do show less nocturnal hypoglycaemia with analogues.
Tiers are applied consistently across the library and re-checked when new trials read out. Read the grading method.
Key studies
The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus Preclinical only
Intensive therapy reduced the development of retinopathy by about 76% and its progression by about 54%, microalbuminuria by 39%, clinical albuminuria by 54% and clinical neuropathy by 60%. Severe hypoglycaemia was roughly threefold more common with intensive therapy. This trial established the glucose hypothesis, and it did so using human insulin. No analogue has ever produced comparable outcome evidence.
Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33) Preclinical only
Intensive control achieved a median HbA1c of 7.0% versus 7.9%, producing a 12% reduction in any diabetes-related endpoint (p=0.029), driven by a 25% reduction in microvascular endpoints (p=0.0099). The reduction in myocardial infarction (16%) narrowly missed significance. Together with the DCCT, this defines what glycaemic control achieves, and the insulin used was human insulin.
Long-acting insulin analogues versus NPH insulin (human isophane insulin) for type 2 diabetes mellitus Preclinical only
No clinically relevant difference in HbA1c between long-acting analogues and isophane insulin. Analogues showed modest reductions in symptomatic and nocturnal hypoglycaemia, but the reviewers found no evidence whatever on patient-relevant long-term outcomes: mortality, morbidity, complications or quality of life.
Association of Initiation of Basal Insulin Analogs vs Neutral Protamine Hagedorn Insulin With Hypoglycemia-Related Emergency Department Visits or Hospital Admissions and With Glycemic Control in Patients With Type 2 Diabetes Preclinical only
Hypoglycaemia-related emergency visits or admissions occurred at 11.9 per 1,000 person-years with analogues versus 8.8 with isophane insulin; the adjusted hazard ratio after matching was 1.16 (95% CI 0.71-1.78). HbA1c improvement was comparable, marginally favouring isophane insulin. In routine practice, analogues did not reduce serious hypoglycaemia, a finding with substantial implications for formulary policy.
The treat-to-target trial: randomized addition of glargine or human NPH insulin to oral therapy of type 2 diabetic patients Preclinical only
Both isophane insulin and glargine achieved a mean HbA1c of about 6.96%, with roughly 60% of participants reaching below 7%. Isophane insulin matched the analogue on efficacy; the analogue's advantage was confined to nocturnal hypoglycaemia, about a quarter less frequent. The clearest single demonstration that isophane insulin, systematically titrated, controls glucose as well as a modern analogue.
Safety
Hypoglycaemia is the dose-limiting adverse effect, and the pronounced four-to-ten-hour peak makes nocturnal hypoglycaemia a particular hazard when a dose is given at bedtime. Inadequate resuspension is a specific and avoidable problem: because the insulin is a crystal suspension, failure to mix gently and completely before each dose delivers an unpredictable amount and is a recognised cause of both hyperglycaemia and hypoglycaemia. The suspension must never be shaken vigorously, which damages the crystals. Protamine carries its own risk: prior exposure to protamine-containing insulin is an established risk factor for severe hypersensitivity (including anaphylaxis and cardiovascular collapse) when protamine sulfate is later given to reverse heparin during cardiac surgery, and this should be flagged before any such procedure. Weight gain, hypokalaemia, injection-site lipohypertrophy from poor site rotation, and local injection reactions occur as with all insulins. Uncommon effects include systemic insulin allergy, lipoatrophy, peripheral oedema, transient refractive change on rapid glycaemic improvement, and anti-insulin antibodies. Renal or hepatic impairment reduces insulin clearance. Concurrent thiazolidinedione therapy increases fluid retention and heart-failure risk. Recurrent hypoglycaemia causes hypoglycaemia unawareness, with implications for driving; UK drivers on insulin must comply with DVLA requirements. The availability of isophane insulin without prescription in most US states is an important safety consideration in its own right: insulin is a high-alert medicine and initiating or adjusting it without clinical supervision has caused avoidable deaths. Non-medical use by people without diabetes has caused deaths and permanent hypoglycaemic brain injury.
Regulatory status
| Jurisdiction | Status |
|---|---|
| United Kingdom | Licensed by the MHRA for the treatment of diabetes mellitus in adults and children. Prescription-only medicine (POM): the US over-the-counter status does not apply in the UK. Recommended in NICE guidance NG28 as the first-choice basal insulin for type 2 diabetes on cost-effectiveness grounds, with analogues reserved for defined circumstances such as recurrent nocturnal hypoglycaemia or the need for assistance with injections. Included on the WHO Model List of Essential Medicines. |
| United States | Human isophane insulin has been marketed for decades and has been regulated as a biological product since March 2020. It is one of very few insulins that can be sold without a prescription in most US states, a genuinely important access route for people who cannot afford analogues or lack insurance, though buying insulin without clinical supervision carries obvious hazards. Available as a stand-alone suspension and in fixed-ratio premixed combinations with soluble or rapid-acting insulin. |
| WADA (sport) | Prohibited at all times, in and out of competition, under section S4 of the WADA Prohibited List, which covers insulins and insulin-mimetics without distinguishing between human insulin and analogues. Athletes with diabetes require a Therapeutic Use Exemption. |
Questions
No, and this is the most common misconception about it. The insulin molecule in isophane insulin is unmodified human insulin, identical in sequence to the hormone the pancreas produces: no substituted residues, no deletions, no fatty acid chains. What is engineered is the formulation: the insulin is co-crystallised with protamine, a basic protein from fish sperm, and zinc. All the prolonged action comes from how slowly those crystals dissolve under the skin.
NPH stands for neutral protamine Hagedorn: neutral pH, protamine as the complexing agent, and Hans Christian Hagedorn, the Danish physician who developed it in 1936. 'Isophane' describes the stoichiometry: protamine and insulin are present in the ratio at which they co-crystallise completely, with neither in excess. Both terms describe the same product.
It is a suspension of microcrystals that settle out of the liquid on standing, so the concentration is not uniform in an unmixed vial or pen. It must be mixed gently (rolling and inverting, never shaking, which damages the crystals) until uniformly cloudy. Incomplete resuspension delivers an unpredictable dose and is a well-documented cause of both unexplained high glucose and unexpected hypoglycaemia.
For HbA1c, the randomised evidence says yes: treat-to-target trials and Cochrane review evidence find no clinically relevant difference. For nocturnal hypoglycaemia, analogues are modestly better, typically around a quarter fewer episodes, because isophane's four-to-ten-hour peak does not match physiological basal secretion. For long-term complications, isophane insulin is the only one with landmark outcome data behind it, from the DCCT and UKPDS. NICE recommends it as first-choice basal insulin in type 2 diabetes on cost-effectiveness grounds.
Because of the protamine. Repeated exposure to protamine-containing insulin can induce antibodies against protamine, and protamine sulfate is the standard agent used to reverse heparin at the end of cardiopulmonary bypass. In sensitised patients this can provoke severe hypersensitivity reactions including anaphylaxis and cardiovascular collapse. Anyone who has used isophane or premixed protamine insulins should make sure the surgical and anaesthetic team knows.