Insulin human with insulin isophane human

What it is

These are the human insulin protein itself, made in bacteria (Lilly) or yeast (Novo Nordisk) by recombinant DNA with a sequence identical to human insulin. Insulin human injection, called regular insulin, acts quickly; insulin isophane human, called NPH, is the same protein crystallised with protamine so it is released slowly. A 70/30 product is one suspension containing 70% isophane and 30% regular, at 100 units per mL, in a 10 mL vial or a 3 mL prefilled pen. The US labels approve the 70/30 mixes to improve glycemic control in adults with diabetes mellitus, and they are dosed in units, never by weight.

What the research says

Insulin lowers blood glucose by driving glucose into muscle and fat and shutting off the liver's glucose output. The outcome evidence is among the strongest in medicine and comes from trials run on human insulin: in type 1 diabetes, intensive treatment cut new retinopathy by 76% (95% CI 62 to 85%) and clinical neuropathy by 60% (95% CI 38 to 74%) over 6.5 years; in newly diagnosed type 2 diabetes it cut microvascular endpoints by 25% (7-40, p=0.0099) over 10 years but did not reduce death. The harm is inseparable from the mechanism: severe hypoglycaemia rose two- to three-fold in the type 1 trial, major hypoglycaemia ran at 1.8% a year on insulin versus 0.7% on conventional care in type 2, and insulin-related hypoglycaemia and dosing errors account for roughly 98,000 US emergency visits a year. Human insulin has also been tested head to head against the newer analogues: Cochrane found no clear benefit of short-acting analogues over regular human insulin, but found more hypoglycaemia with isophane (NPH) than with long-acting analogues, with an absolute difference of about one person in a hundred. A fixed 70/30 mix is less flexible than separate insulins, and in one 371-person trial twice-daily human 70/30 gave worse control and more than twice the hypoglycaemia of a once-daily analogue added to tablets.

Evidence grade: Well established.

How it works

Drug class: Insulin (human insulin and human isophane/NPH insulin; a short-acting and an intermediate-acting insulin supplied as a fixed 70/30 premixed suspension)

Insulin tells muscle and fat cells to pull glucose out of the blood, and tells the liver to stop making and releasing glucose. It also stops the body breaking down fat and protein for fuel and pushes it to build protein instead. In the 70/30 mix the 30% regular insulin works within the hour for the meal just eaten, while the 70% isophane is released slowly over most of a day to cover the background need. (Source 1)

What it is used for

• The trial that established the value of tight control in type 1 diabetes was run entirely on human insulin, and found large reductions in retinopathy, nephropathy and neuropathy over 6.5 years at the cost of two to three times as much severe hypoglycaemia. Evidence: established. (Source 2)
• UKPDS 33 showed intensive control with insulin or a sulfonylurea cut microvascular complications by 25% over 10 years but did not significantly reduce death; ACCORD later showed that pushing HbA1c toward normal in high-risk patients increased mortality, so the benefit depends on how far control is pushed and in whom. Evidence: established. (Source 3)
• In a 371-person randomised comparison, human 70/30 twice daily achieved less HbA1c reduction and more than twice the rate of confirmed hypoglycaemia (9.87 versus 4.07 episodes per patient-year) compared with adding once-daily insulin glargine to oral agents. The trial was open-label and the regimens differed in more than the insulin. Evidence: disputed. (Source 4)
• Cochrane found no clear benefit of short-acting analogues over regular human insulin in type 2 diabetes (HbA1c difference -0.03%, 95% CI -0.16 to 0.09), while long-acting analogues did reduce hypoglycaemia against isophane insulin, with an absolute benefit of roughly one person in a hundred. Evidence: established. (Source 5)
• Regular insulin human is used intravenously in hospital for these purposes and a ready-to-use intravenous human insulin product (Myxredlin) exists, but we did not retrieve outcome-trial evidence for these specific uses in this run, and the 70/30 suspension is not for intravenous use. We cannot characterise the evidence from what we read. Evidence: unknown. (Source 6)

Interactions

• Alcohol (label): Alcohol can either lower or raise the blood-glucose-lowering effect of insulin, so it can cause hypoglycaemia or loss of control in either direction. The label groups it with beta-blockers, clonidine and lithium salts and says dose adjustment and more frequent glucose monitoring may be needed. (Source 7)
• Food (eating less than the dose was planned for, or changing meal timing) (clinical trial): Because the dose is fixed to an expected meal, eating less or later than planned is a direct cause of hypoglycaemia. In US national surveillance it was one of the two commonest identified triggers of insulin emergencies, and the label lists changes in meal pattern, including macronutrient content and timing, among the factors that raise hypoglycaemia risk. (Source 8)
• Beta-blockers, clonidine, guanethidine and reserpine (label): These can mask the warning signs of a low blood sugar, so a hypoglycaemic episode may go unnoticed until it is severe. The label says more frequent glucose monitoring may be needed. (Source 7)
• Salicylates (including aspirin), fluoxetine, fibrates, ACE inhibitors and other glucose-lowering drugs (label): These increase the risk of hypoglycaemia when combined with insulin, so the insulin dose may need lowering and glucose checking more often. (Source 9)
• Niacin (nicotinic acid) and thyroid hormones (label): Niacin and thyroid hormones are both named in the label's row of drugs that may DECREASE insulin's blood-glucose-lowering effect, meaning blood glucose may run higher and more insulin may be needed. The label says dose adjustment and more frequent glucose monitoring may be required. Niacin is also sold as a supplement, which is why it matters here. This is the opposite direction from the drugs that raise the risk of hypoglycaemia, which are a separate row of the same table. (Source 7)
• Thiazolidinediones (pioglitazone, rosiglitazone) (label): Combined with insulin these can cause dose-related fluid retention, which may lead to or worsen heart failure. The label says people on both should be watched for signs of heart failure and that stopping or reducing the thiazolidinedione must be considered if it develops. (Source 10)
• Potassium-lowering drugs and drugs sensitive to serum potassium (label): Insulin shifts potassium into cells and can lower blood potassium, so the label says potassium should be monitored in people already at risk, such as those on potassium-lowering medicines. (Source 11)
• St John's wort, grapefruit, calcium, iron, magnesium, fish oil, turmeric, red yeast rice (label): We found no documented interaction between human insulin and any of these in the sources we read. Insulin is a protein given by injection and is not metabolised by the cytochrome P450 enzymes that grapefruit and St John's wort act on, so the usual mechanism for those interactions does not apply. The label's complete drug-interaction table is quoted here in full for that reason: across all four of its rows the only substance also sold as a supplement is niacin, which is covered separately above. (Source 7)

Stopping it

• Human insulin causes no dependence or withdrawal syndrome in the pharmacological sense, and neither label we read has a drug abuse and dependence section claiming one. What stopping means depends on the person: someone with type 1 diabetes who stops insulin develops life-threatening hyperglycaemia, while in type 2 diabetes it may be reduced or stopped if other treatment or weight change makes it unnecessary. Either way it is a prescriber's decision, and the label frames every regimen change as a reason to monitor glucose more closely. (Source 12)
• Changing between insulin products is the risky part rather than stopping as such: in US surveillance, administration of the wrong insulin product was one of the two commonest triggers of emergency visits for insulin-related harm, alongside eating less than planned. (Source 13)
• One site-specific reason for a sudden change in effect on stopping or switching injection sites is skin amyloid: the label records that hyperglycaemia has been reported with repeated injections into areas of localized cutaneous amyloidosis and hypoglycaemia with a sudden change to an unaffected site. This sits in the label's post-marketing list, which it introduces by saying the reactions are reported voluntarily from a population of uncertain size, so neither their frequency nor their causal link to the drug is established. (Source 14)
• If treatment is stopped because of too much insulin rather than too little, the label's position is that mild hypoglycaemia is treated with oral glucose, severe episodes with glucagon or intravenous glucose, and that observation must continue because hypoglycaemia can recur after apparent recovery. (Source 15)

What goes wrong

The same trial found the price of intensive insulin treatment was two to three times as much severe hypoglycaemia. (Source 2)

• Randomized trial, High certainty.
• Size: 1,441 patients.
• Who: people with insulin-dependent diabetes mellitus.
• How long: mean 6.5 years.
• Result: a two-to-threefold increase in severe hypoglycemia; the abstract gives no absolute rate.
• Funding: publicly funded (Research Support, U.S. Gov't, P.H.S.)

The chief adverse event associated with intensive therapy was a two-to-threefold increase in severe hypoglycemia.

In the same trial, insulin caused more severe hypoglycaemia and more weight gain than either the sulfonylureas or conventional care. (Source 16)

• Randomized trial, High certainty.
• Size: 3,867 patients, analysed both by intention to treat and by actual therapy.
• Who: newly diagnosed type 2 diabetes.
• How long: over 10 years.
• Result: major hypoglycaemic episodes per year: 0.7% conventional, 1.0% chlorpropamide, 1.4% glibenclamide, 1.8% insulin. Weight gain: mean 2.9 kg in the intensive group overall, 4.0 kg on insulin versus 2.6 kg chlorpropamide and 1.7 kg glibenclamide.
• Funding: publicly and charitably funded. Collaborative group authorship.

The rates of major hypoglycaemic episodes per year were 0.7% with conventional treatment, 1.0% with chlorpropamide, 1.4% with glibenclamide, and 1.8% with insulin. Weight gain was significantly higher in the intensive group (mean 2.9 kg) than in the conventional group (p<0.001), and patients assigned insulin had a greater gain in weight (4.0 kg) than those assigned chlorpropamide (2.6 kg) or glibenclamide (1.7 kg).

Pushing glucose to near-normal in high-risk type 2 diabetes increased death rather than preventing heart attacks, which bounds how aggressively insulin should be used. (Source 17)

• Randomized trial, High certainty.
• Size: 10,251 patients (mean age 62.2; 35% with a previous cardiovascular event)
• Who: people with type 2 diabetes and established cardiovascular disease or additional risk factors, median HbA1c 8.1% at entry.
• How long: mean 3.5 years, when intensive therapy was stopped early for excess mortality.
• Result: deaths 257 versus 203 (hazard ratio 1.22; 95% CI 1.01 to 1.46; P=0.04), an absolute excess of 54 deaths; primary composite outcome 352 versus 371 (hazard ratio 0.90; 95% CI 0.78 to 1.04; P=0.16). Hypoglycaemia requiring assistance and weight gain over 10 kg were both more frequent on intensive therapy (P<0.001)
• Funding: publicly funded (Research Support, N.I.H., Extramural). Published by the ACCORD Study Group with no named individual first author. The intensive arm used multiple drugs including insulin, so the harm is attributable to the strategy rather than to insulin alone.

At the same time, 257 patients in the intensive-therapy group died, as compared with 203 patients in the standard-therapy group (hazard ratio, 1.22; 95% CI, 1.01 to 1.46; P=0.04). Hypoglycemia requiring assistance and weight gain of more than 10 kg were more frequent in the intensive-therapy group (P<0.001).

Against the long-acting analogues glargine and detemir, human isophane (NPH) insulin given on its own was linked to more hypoglycaemia, but only the detemir comparison for serious hypoglycaemia reached statistical significance, and the absolute difference was about one person in a hundred. (Source 18)

• Systematic review, Very low certainty.
• Size: 24 RCTs; 3,419 people randomised to glargine and 1,321 to detemir, against NPH comparators (6,164 participants in the glargine severe-hypoglycaemia analysis)
• Who: adults with type 2 diabetes in trials targeting near-normal blood glucose, given NPH insulin as NPH rather than as part of a premixed insulin.
• How long: 24 weeks to five years.
• Result: severe hypoglycaemia, glargine versus NPH: RR 0.68 (95% CI 0.46 to 1.01; P = 0.06; absolute risk reduction -1.2%, 95% CI -2.0 to 0; very low certainty). Serious hypoglycaemia, detemir versus NPH: Peto odds ratio 0.16 (95% CI 0.04 to 0.61; P = 0.007; ARR -0.9%, 95% CI -1.1 to -0.4; low certainty). HbA1c was comparable.
• Funding: not stated; the record lists Research Support, Non-U.S. Gov't. All trials had unclear or high risk of bias in several domains, and the glucose targets used were lower than current practice.

Limit of this finding: This review compared the analogues with NPH insulin given on its own, not with the isophane component of a 70/30 premix, so it does not test HUMULIN 70/30 and says nothing about the soluble insulin in the mix. Two of the three figures here are also not statistically significant on the review's own terms: severe hypoglycaemia with glargine versus NPH, RR 0.68 (95% CI 0.46 to 1.01; P = 0.06), and with detemir, RR 0.45 (95% CI 0.17 to 1.20; P = 0.11), both have intervals that include 1. Only the Peto odds ratio for serious hypoglycaemia with detemir (0.16, 95% CI 0.04 to 0.61; P = 0.007) excludes it. The review also rates the severe-hypoglycaemia evidence very low certainty, and the abstract prints that Peto odds ratio with an unmatched closing bracket, which is the source's typesetting and not a transcription error. Read this as a hint of fewer low-sugar episodes on the analogues rather than a demonstrated difference.

Insulin glargine compared to NPH insulin had a risk ratio (RR) for severe hypoglycaemia of 0.68 (95% confidence interval (CI) 0.46 to 1.01; P = 0.06; absolute risk reduction (ARR) -1.2%, 95% CI -2.0 to 0; 14 trials, 6164 participants; very low-certainty evidence). The RR for serious hypoglycaemia was 0.75 (95% CI 0.52 to 1.09; P = 0.13; ARR -0.7%, 95% CI -1.3 to 0.2; 10 trials, 4685 participants; low-certainty evidence). Treatment with insulin glargine reduced the incidence of confirmed hypoglycaemia and confirmed nocturnal hypoglycaemia. Treatment with insulin detemir compared to NPH insulin found an RR for severe hypoglycaemia of 0.45 (95% CI 0.17 to 1.20; P = 0.11; ARR -0.9%, 95% CI -1.4 to 0.4; 5 trials, 1804 participants; very low-certainty evidence). The Peto odds ratio for serious hypoglycaemia was 0.16, 95% CI 0.04 to 0.61; P = 0.007; ARR -0.9%, 95% CI -1.1 to -0.4; 5 trials, 1777 participants; low-certainty evidence).

Cochrane put the absolute size of that hypoglycaemia advantage over human isophane insulin at about one person in a hundred. (Source 19)

• Systematic review, Low certainty.
• Size: 24 RCTs, 4,740 people on analogues plus NPH comparators.
• Who: adults with type 2 diabetes.
• How long: 24 weeks to five years.
• Result: serious hypoglycaemic events were rare and the absolute risk reducing effect was low; approximately one in 100 people treated with insulin detemir instead of NPH benefited.
• Funding: not stated; the record lists Research Support, Non-U.S. Gov't.

Limit of this finding: The review attaches an applicability limit to this conclusion that matters more than the figure: the trials set "low blood glucose and HbA1c targets, corresponding to near normal or even non-diabetic blood glucose levels", while current guidelines recommend less-intensive lowering for most people with type 2 diabetes in practice. The review says it therefore remains unclear whether the same effects would be seen in daily clinical practice, and that most trials did not report patient-relevant outcomes at all.

However, serious hypoglycaemic events were rare and the absolute risk reducing effect was low. Approximately one in 100 people treated with insulin detemir instead of NPH insulin benefited.

In a head-to-head trial starting insulin in type 2 diabetes, twice-daily human 70/30 controlled glucose less well than a once-daily analogue added to tablets, and caused more than twice as many hypoglycaemic episodes. (Source 4)

• Randomized trial, Moderate certainty.
• Size: 371 insulin-naive patients.
• Who: insulin-naive people with type 2 diabetes poorly controlled on sulfonylurea plus metformin (fasting blood glucose 120 mg/dl or more, HbA1c 7.5-10.5%)
• How long: 24 weeks.
• Result: confirmed hypoglycaemic episodes 9.87 versus 4.07 per patient-year on 70/30 versus glargine plus oral agents (P < 0.0001); HbA1c fall -1.31% versus -1.64% (P = 0.0003); reaching HbA1c 7.0% or less without confirmed nocturnal hypoglycaemia 28.6% versus 45.5% (P = 0.0013)
• Funding: not stated in the abstract; the record lists Research Support, Non-U.S. Gov't. The trial was open-label, not blinded, which matters for a subjective outcome like hypoglycaemia, and it compared different regimens as well as different insulins (the 70/30 arm stopped its oral agents).

Glargine plus OAD patients had fewer confirmed hypoglycemic episodes than 70/30 patients (mean 4.07 vs. 9.87/patient-year, P < 0.0001).

Insulin-related hypoglycaemia and dosing errors account for roughly 98,000 US emergency department visits a year, about a third of which end in hospital admission. (Source 13)

• Survey study, Moderate certainty.
• Size: 8,100 surveillance cases, projected nationally; denominators from a national household survey of insulin use.
• Who: insulin-treated people with diabetes attending US emergency departments, 2007-2011.
• How long: five years of surveillance.
• Result: an estimated 97,648 (95% CI 64,410-130,887) ED visits annually; 29.3% (95% CI 21.8%-36.8%) resulted in hospitalisation; severe neurologic sequelae in 60.6% (95% CI 51.3%-69.9%); blood glucose 50 mg/dL or less in 53.4%. People 80 or older were 2.5 times as likely to attend (95% CI 1.5-4.3) and 4.9 times as likely to be hospitalised (95% CI 2.6-9.1) as those aged 45-64.
• Funding: US public health surveillance (CDC project). Observational and projected from a sample, so the national totals are estimates with wide intervals.

Limit of this finding: These are national estimates projected from a sample of 8,100 surveillance cases collected between 1 January 2007 and 31 December 2011, with denominators from a household survey of insulin use. The intervals are wide, the design is surveillance projected to the population rather than a survey of individuals, and the figures are now more than a decade old; insulin products and glucose monitoring have changed since.

an estimated 97,648 (95% CI, 64,410-130,887) ED visits for IHEs occurred annually; almost one-third (29.3%; 95% CI, 21.8%-36.8%) resulted in hospitalization.

The commonest triggers for those emergencies were eating less than planned and injecting the wrong insulin product. (Source 13)

• Survey study, Moderate certainty.
• Size: 8,100 surveillance cases.
• Who: insulin-treated people attending US emergency departments, 2007-2011.
• How long: five years.
• Result: no rates given for individual precipitants; the two most commonly identified were reduced food intake and administration of the wrong insulin product.
• Funding: US public health surveillance (CDC project)

Limit of this finding: These are national estimates projected from a sample of 8,100 surveillance cases collected between 1 January 2007 and 31 December 2011, with denominators from a household survey of insulin use. The intervals are wide, the design is surveillance projected to the population rather than a survey of individuals, and the figures are now more than a decade old; insulin products and glucose monitoring have changed since.

The most commonly identified IHE precipitants were reduced food intake and administration of the wrong insulin product.

Low blood potassium is a documented harm of all insulins because insulin drives potassium into cells. (Source 11)

• Official position, Low certainty.
• Size: not stated.
• Who: anyone receiving insulin, particularly people on potassium-lowering drugs.
• How long: not applicable.
• Result: no rate is given; the label states untreated hypokalemia may cause respiratory paralysis, ventricular arrhythmia and death.
• Funding: manufacturer's label (Eli Lilly), a regulatory position dated 2026-06-17.

All insulins, including HUMULIN 70/30, cause a shift in potassium from the extracellular to intracellular space, possibly leading to hypokalemia. Untreated hypokalemia may cause respiratory paralysis, ventricular arrhythmia, and death.

Injecting insulin under the skin has caused loss of fat (lipoatrophy) or thickening of the tissue (lipohypertrophy) at injection sites. (Source 14)

• Official position, Low certainty.
• Size: not stated; postmarketing reports.
• Who: people injecting insulin subcutaneously.
• How long: with repeated injections over time.
• Result: no rates given; the label records lipoatrophy and lipohypertrophy as reactions that have occurred in some patients.
• Funding: manufacturer's label (Eli Lilly), a regulatory position dated 2026-06-17.

Limit of this finding: These are post-marketing reports, and the label itself says so immediately above the list: they are "reported voluntarily from a population of uncertain size", so "it is not always possible to reliably estimate their frequency or to establish a causal relationship to drug exposure". Treat them as things that have been reported with insulin rather than as rates or as established effects.

Administration of insulin subcutaneously, including HUMULIN 70/30, has resulted in lipoatrophy (depression in the skin) or lipohypertrophy (enlargement or thickening of tissue) [see Dosage and Administration (2.1)] in some patients.

Localized cutaneous amyloidosis has occurred at injection sites, and the label records high blood sugar when insulin is repeatedly injected into such an area and low blood sugar on moving suddenly to an unaffected site. (Source 14)

• Official position, Low certainty.
• Size: not stated; postmarketing reports.
• Who: people injecting insulin subcutaneously.
• How long: with repeated injections over time.
• Result: no rates given. The label records that hyperglycemia has been reported with repeated insulin injections into areas of localized cutaneous amyloidosis, and hypoglycemia with a sudden change to an unaffected site.
• Funding: manufacturer's label (Eli Lilly), a regulatory position dated 2026-06-17.

Limit of this finding: These are post-marketing reports, and the label itself says so immediately above the list: they are "reported voluntarily from a population of uncertain size", so "it is not always possible to reliably estimate their frequency or to establish a causal relationship to drug exposure". Treat them as things that have been reported with insulin rather than as rates or as established effects.

Localized cutaneous amyloidosis at the injection site has occurred. Hyperglycemia has been reported with repeated insulin injections into areas of localized cutaneous amyloidosis; hypoglycemia has been reported with a sudden change to an unaffected injection site.

Weight gain is a recognised effect of insulin treatment, attributed to insulin's anabolic action and to glucose no longer being lost in the urine. (Source 14)

• Official position, Low certainty.
• Size: not stated in the label; UKPDS 33 measured 4.0 kg mean gain on insulin over 10 years.
• Who: people treated with insulin.
• How long: not stated in the label.
• Result: the label gives no figure; UKPDS 33 recorded a mean 4.0 kg gain on insulin against 2.6 kg on chlorpropamide and 1.7 kg on glibenclamide.
• Funding: manufacturer's label (Eli Lilly), a regulatory position dated 2026-06-17.

Limit of this finding: These are post-marketing reports, and the label itself says so immediately above the list: they are "reported voluntarily from a population of uncertain size", so "it is not always possible to reliably estimate their frequency or to establish a causal relationship to drug exposure". Treat them as things that have been reported with insulin rather than as rates or as established effects.

Weight gain has occurred with insulins, including HUMULIN 70/30, and has been attributed to the anabolic effects of insulin and the decrease in glycosuria.

Inhaled human insulin carries a boxed warning for acute bronchospasm that the injectable human insulins do not. (Source 20)

• Official position, Moderate certainty.
• Size: not stated in the boxed warning.
• Who: people with asthma or chronic obstructive pulmonary disease given inhaled insulin human (AFREZZA)
• How long: not applicable.
• Result: no rate in the boxed warning; the product is contraindicated in chronic lung disease and spirometry is required before starting.
• Funding: manufacturer's label (MannKind), a regulatory position dated 2026-05-30. This applies to the inhaled formulation only; we parsed LOINC 34066-1 across Humulin 70/30, Novolin 70/30, Humulin R, Humulin R U-500, Humulin N, Novolin N and Myxredlin and found no boxed warning in any of them.

Acute bronchospasm has been observed in AFREZZA-treated patients with asthma and Chronic Obstructive Pulmonary Disease (COPD).

What the evidence supports

In type 1 diabetes, intensive human-insulin treatment aimed at near-normal glucose sharply cut the development and progression of eye, kidney and nerve complications. (Source 2)

• Randomized trial, High certainty.
• Size: 1,441 patients (726 with no retinopathy at baseline, 715 with mild retinopathy)
• Who: people with insulin-dependent diabetes mellitus; human insulin was the only insulin available when the trial ran.
• How long: mean 6.5 years of follow-up.
• Result: development of retinopathy reduced 76% (95% CI 62 to 85%) in the primary-prevention cohort; progression slowed 54% (95% CI 39 to 66%); microalbuminuria reduced 39% (95% CI 21 to 52%); albuminuria 54% (95% CI 19 to 74%); clinical neuropathy 60% (95% CI 38 to 74%). These are relative reductions; the paper reports adjusted mean risk reductions rather than absolute event differences in the abstract.
• Funding: publicly funded (Research Support, U.S. Gov't, P.H.S.). Published by a collaborative group; the paper has no individual named first author.

In the primary-prevention cohort, intensive therapy reduced the adjusted mean risk for the development of retinopathy by 76 percent (95 percent confidence interval, 62 to 85 percent), as compared with conventional therapy. In the secondary-intervention cohort, intensive therapy slowed the progression of retinopathy by 54 percent (95 percent confidence interval, 39 to 66 percent) and reduced the development of proliferative or severe nonproliferative retinopathy by 47 percent (95 percent confidence interval, 14 to 67 percent). In the two cohorts combined, intensive therapy reduced the occurrence of microalbuminuria (urinary albumin excretion of > or = 40 mg per 24 hours) by 39 percent (95 percent confidence interval, 21 to 52 percent), that of albuminuria (urinary albumin excretion of > or = 300 mg per 24 hours) by 54 percent (95 percent confidence interval 19 to 74 percent), and that of clinical neuropathy by 60 percent (95 percent confidence interval, 38 to 74 percent).

What the evidence does not support

Cochrane found that the newer short-acting insulin analogues give no clear advantage over plain regular human insulin in type 2 diabetes. (Source 5)

• Systematic review, Low certainty.
• Size: 2,751 participants across 10 trials (1,388 analogue, 1,363 regular human insulin)
• Who: adult, non-pregnant people with type 2 diabetes.
• How long: 24 to 104 weeks, mean about 41 weeks.
• Result: HbA1c mean difference -0.03% (95% CI -0.16 to 0.09; P = 0.60; 9 trials, 2608 participants; low certainty); deaths 5/1272 (0.4%) on analogues versus 3/1247 (0.2%) on human insulin (Peto OR 1.66, 95% CI 0.41 to 6.64); non-severe hypoglycaemic episodes per participant per month mean difference 0.08 events (95% CI 0.00 to 0.16; P = 0.05; very low certainty)
• Funding: not stated; the record lists Research Support, Non-U.S. Gov't. None of the trials was blinded and nine of ten were at high risk of performance and detection bias for subjective outcomes like hypoglycaemia.

The MD in glycosylated haemoglobin A1c (HbA1c) change was -0.03% (95% CI -0.16 to 0.09; P = 0.60; 9 trials, 2608 participants; low-certainty evidence).

Cochrane's own summary was that no clear benefit of short-acting analogues over regular human insulin could be demonstrated, and the certainty of evidence was poor. (Source 21)

• Systematic review, Low certainty.
• Size: 2,751 participants across 10 trials.
• Who: adult, non-pregnant people with type 2 diabetes.
• How long: 24 to 104 weeks.
• Result: no advantage demonstrated on mortality, complications or severe hypoglycaemia; the review reports these outcomes as sparse.
• Funding: not stated; the record lists Research Support, Non-U.S. Gov't.

Our analysis found no clear benefits of short-acting insulin analogues over regular human insulin in people with type 2 diabetes.

In the same trial, fewer people on human 70/30 reached the fasting glucose target than on the analogue regimen. (Source 4)

• Randomized trial, Moderate certainty.
• Size: 371 insulin-naive patients.
• Who: insulin-naive people with type 2 diabetes on sulfonylurea plus metformin.
• How long: 24 weeks, with weekly forced titration.
• Result: reached target fasting blood glucose of 100 mg/dl or less: 15.0% on 70/30 versus 31.6% on glargine plus oral agents (P = 0.0001); adjusted mean fasting glucose difference -17 mg/dl (-0.9 mmol/l), P < 0.0001.
• Funding: not stated in the abstract; open-label design.

more patients reached target FBG </=100 mg/dl with glargine plus OAD than with 70/30 (31.6 vs. 15.0%, P = 0.0001)

A fixed 70/30 mix removes the ability to adjust the basal and mealtime components separately, which is a structural limitation rather than a side effect. (Source 12)

• Official position, Low certainty.
• Size: not applicable.
• Who: anyone using a premixed insulin.
• How long: not applicable.
• Result: no numeric effect; the label states independent adjustment of the basal or prandial dose is not possible with a premixed insulin.
• Funding: manufacturer's label (Eli Lilly), a regulatory position dated 2026-06-17.

The proportion of rapid acting and long acting insulin is fixed in a premixed insulin such as HUMULIN 70/30. Independent adjustment of the basal or prandial dose is not possible when using a premixed insulin.

The same Cochrane review found no meaningful difference between the long-acting analogues and human isophane insulin on deaths, diabetes complications, quality of life, weight gain or the overall rate of adverse events. (Source 18)

• Systematic review, Low certainty.
• Size: 24 randomised trials; 3,419 people randomised to insulin glargine and 1,321 to insulin detemir.
• Who: adults with type 2 diabetes mellitus.
• How long: 24 weeks to five years.
• Result: information on death from any cause, diabetes-related complications, health-related quality of life and socioeconomic effects was insufficient or lacking in almost all trials; where data existed there were no meaningful differences; no clear difference in weight gain; comparable incidence of adverse events. No trial at all compared insulin glargine U300 or insulin degludec with NPH insulin.
• Funding: Cochrane review; trial-level funding not stated in the abstract.

Limit of this finding: "No meaningful difference" here mostly means too little data to tell: the review says the information was insufficient or lacking in almost all the included trials, and all trials had unclear or high risk of bias in several domains. This is an absence of evidence of a difference, not evidence that the insulins are equivalent on these outcomes.

Information on patient-relevant outcomes such as death from any cause, diabetes-related complications, health-related quality of life and socioeconomic effects was insufficient or lacking in almost all included trials. For those outcomes for which some data were available, there were no meaningful differences between treatment with glargine or detemir and treatment with NPH.

Where the evidence is mixed

In newly diagnosed type 2 diabetes, intensive glucose control with insulin or a sulfonylurea cut microvascular complications by about a quarter but did not reduce death. (Source 3)

• Randomized trial, High certainty.
• Size: 3,867 newly diagnosed patients with type 2 diabetes.
• Who: newly diagnosed type 2 diabetes, median age 54, with fasting plasma glucose 6.1-15.0 mmol/L after 3 months of diet.
• How long: over 10 years.
• Result: HbA1c 7.0% versus 7.9%; any diabetes-related endpoint 12% lower (95% CI 1-21, p=0.029); diabetes-related death 10% lower (-11 to 27, p=0.34); all-cause mortality 6% lower (-10 to 20, p=0.44); microvascular endpoints 25% lower (7-40, p=0.0099). No difference between insulin and the two sulfonylureas.
• Funding: publicly and charitably funded (Research Support, Non-U.S. Gov't and U.S. Gov't, P.H.S.). Published by the UK Prospective Diabetes Study (UKPDS) Group with no named individual first author.

Compared with the conventional group, the risk in the intensive group was 12% lower (95% CI 1-21, p=0.029) for any diabetes-related endpoint; 10% lower (-11 to 27, p=0.34) for any diabetes-related death; and 6% lower (-10 to 20, p=0.44) for all-cause mortality. Most of the risk reduction in the any diabetes-related aggregate endpoint was due to a 25% risk reduction (7-40, p=0.0099) in microvascular endpoints, including the need for retinal photocoagulation.

Where the research disagrees

Whether human insulin is adequate or whether insulin analogues are needed

• Fullerton and colleagues, Cochrane review of 10 trials in 2,751 adults with type 2 diabetes, systematic-review: Our analysis found no clear benefits of short-acting insulin analogues over regular human insulin in people with type 2 diabetes. (Source 21)
• Semlitsch and colleagues, Cochrane review of 24 trials comparing long-acting analogues with human isophane (NPH) insulin, which adds that the trials aimed at near-normal glucose targets and that it is unclear whether the same effects hold in ordinary practice, systematic-review: While the effects on HbA1c were comparable, treatment with insulin glargine and insulin detemir resulted in fewer participants experiencing hypoglycaemia when compared with NPH insulin. (Source 19)

Whether a boxed warning applies to insulin human

• The injectable human insulin labels (Humulin 70/30, Novolin 70/30, Humulin R, Humulin R U-500, Humulin N, Novolin N, Myxredlin), parsed for LOINC 34066-1, position: HUMULIN 70/30 (insulin isophane human and insulin human) injectable suspension is a mixture of 70% insulin isophane human suspension, an intermediate-acting insulin, and 30% insulin human injection, a short-acting insulin. (Source 6)
• The AFREZZA label for inhaled insulin human, which does carry a boxed warning, position: WARNING: RISK OF ACUTE BRONCHOSPASM IN PATIENTS WITH CHRONIC LUNG DISEASE See full prescribing information for complete boxed warning. (Source 20)

How hard to push glucose down with insulin

• The UK Prospective Diabetes Study (UKPDS) Group, from 3,867 newly diagnosed patients over 10 years, rct: Most of the risk reduction in the any diabetes-related aggregate endpoint was due to a 25% risk reduction (7-40, p=0.0099) in microvascular endpoints, including the need for retinal photocoagulation. (Source 3)
• The Action to Control Cardiovascular Risk in Diabetes (ACCORD) Study Group, from 10,251 high-risk patients, rct: At the same time, 257 patients in the intensive-therapy group died, as compared with 203 patients in the standard-therapy group (hazard ratio, 1.22; 95% CI, 1.01 to 1.46; P=0.04). (Source 17)

How much

• Reference intake: There is no reference intake for insulin; the dose is set by the prescriber and titrated to the individual. The FDA-approved label's position is that HUMULIN 70/30 is injected subcutaneously about 30 to 45 minutes before a meal and individualised against the person's metabolic needs, blood glucose monitoring results and glycemic control goal. Insulin is measured in units, not milligrams, and these products supply 100 units per mL. (Source 12)
• Upper limit: The labels set no maximum dose, because requirement varies enormously between people and over time in the same person. The label instead states that dose requirements may change with physical activity, meal patterns, major illness and other drugs, and that monitoring should be increased when the regimen changes. An excess dose is defined by its consequence rather than a number: excess insulin administration may cause hypoglycemia and hypokalemia. (Source 15)
• Studied: The type 1 diabetes trial compared intensive therapy by external insulin pump or three or more daily injections against conventional therapy of one or two daily injections, with no fixed dose specified. (Source 22)
• Studied: Janka 2005 titrated 30% regular / 70% human NPH insulin (70/30) twice daily against once-daily insulin glargine plus glimepiride and metformin, both to a fasting blood glucose target of 100 mg/dl or less, using a weekly forced-titration algorithm. (Source 23)
• Studied: Cochrane's analogue-versus-human-insulin review pooled 10 trials of 24 to 104 weeks in 2,751 adults with type 2 diabetes, 1,363 of them on regular human insulin. (Source 5)
• Studied: In healthy males (n=18) the label reports a 0.3 unit/kg subcutaneous dose of HUMULIN 70/30 beginning to act at about 50 minutes, peaking at about 3.5 hours and lasting about 23 hours. (Source 24)

A common belief, and what the research shows

The belief: Human insulin is the old, inferior product and the modern analogues are simply better.

What the research shows: On the evidence the gap is narrow and uneven. For mealtime insulin, Cochrane verdict was that “Our analysis found no clear benefits of short-acting insulin analogues over regular human insulin in people with type 2 diabetes.”, with an HbA1c difference of “-0.03% (95% CI -0.16 to 0.09; P = 0.60; 9 trials, 2608 participants; low-certainty evidence)”. For background insulin the analogues do reduce hypoglycaemia against human isophane, but the size is small: “However, serious hypoglycaemic events were rare and the absolute risk reducing effect was low. Approximately one in 100 people treated with insulin detemir instead of NPH insulin benefited.” Where human insulin does come off worse is as a fixed twice-daily 70/30 regimen: “Glargine plus OAD patients had fewer confirmed hypoglycemic episodes than 70/30 patients (mean 4.07 vs. 9.87/patient-year, P < 0.0001).” That is a point about the regimen inflexibility as much as about the molecule, since “Independent adjustment of the basal or prandial dose is not possible when using a premixed insulin.”

Questions and answers

What is it?

Insulin human is the human insulin protein made in bacteria or yeast by recombinant DNA, with an amino acid sequence identical to the insulin a human pancreas makes. Insulin isophane human, also called NPH, is the same protein crystallised with protamine so it is absorbed slowly. The 70/30 products are a suspension of 70% isophane and 30% plain insulin human in one vial or pen, at 100 units per mL. (Source 6)

What does it do in the body?

Insulin lowers blood glucose by making muscle and fat take glucose up and by stopping the liver releasing it. It also blocks the breakdown of fat and protein and pushes protein synthesis. A 70/30 mix gives a quick rise in activity from the plain insulin and a long tail from the isophane: in healthy men the effect started at about 50 minutes, peaked at about 3.5 hours and lasted around 23 hours. (Source 1)

Is it good or bad for you?

For someone who cannot make enough insulin it is life-saving, and the long trials show real benefit: in type 1 diabetes intensive human insulin cut new retinopathy by 76% and neuropathy by 60%, and in newly diagnosed type 2 diabetes intensive control cut microvascular complications by 25%. The harm runs in the same direction as the benefit: more insulin means more hypoglycaemia, two to three times as much in the type 1 trial, and ACCORD showed that pushing glucose too low in high-risk type 2 diabetes increased deaths. So the benefit and the danger both come from the same property, and the right amount is a clinical judgement, not a fixed number. (Source 2)

How do you get more of it?

Human insulin is only obtained as an injected medicine, dosed in units and never in milligrams. The 70/30 products are 100 units per mL and are injected under the skin about 30 to 45 minutes before a meal; the amount is individualised by the prescriber against the person's metabolic needs, glucose readings and target. No food or supplement contains it, and swallowing it would destroy it. (Source 12)

If it is harmful, what reduces it?

There is no way to remove injected insulin from the body, which is why an excess is managed by giving glucose rather than by reversing the drug. The label's position is that mild hypoglycaemia can be treated with oral glucose and that more severe episodes with coma, seizure or neurologic impairment may need glucagon or intravenous glucose, with sustained carbohydrate and observation afterwards because hypoglycaemia can come back. Potassium also has to be corrected. (Source 15)

Why might someone be low in it or missing it?

A person can be short of their own insulin because the beta cells that make it have been destroyed, as in type 1 diabetes, or because the body has become resistant to it and the pancreas cannot keep up, as in type 2. For the injected medicine, the practical reasons the dose stops matching need are changes in activity, changes in how much or when someone eats, major illness, changes in kidney or liver function, and other drugs. Kidney or liver impairment also raises hypoglycaemia risk. (Source 8)

Which whole foods contain it or feed it?

No food contains injectable insulin. Food matters in the opposite direction: because the dose is fixed to an expected meal, eating less than planned is one of the two commonest causes of insulin emergencies in US national surveillance data, alongside injecting the wrong insulin product. The label says dose requirements change with meal patterns, meaning both the amount and the timing of what is eaten. (Source 13)

What happens if you do not have it?

Without enough insulin, blood glucose rises and over years damages small blood vessels and nerves. The size of that is quantified by what happened when treatment was intensified instead: new retinopathy fell 76%, microalbuminuria 39% and clinical neuropathy 60% against conventional treatment over 6.5 years. In type 2 diabetes, better control cut microvascular endpoints 25% over 10 years. Acute, complete lack of insulin in type 1 diabetes is a medical emergency. (Source 3)

How can you test for it?

Insulin treatment is judged by blood glucose and by HbA1c, the measure every one of these trials used as its glycaemic endpoint. Self-monitoring of blood glucose is what the label calls essential to preventing and managing hypoglycaemia, with more frequent testing in people at higher risk or with reduced awareness of low glucose. Measuring insulin itself in blood is not a routine test for people on insulin treatment, because injected insulin cannot be distinguished from the body's own in ordinary assays and none of the sources we read describes such a test being used for this purpose. (Source 8)

References

1. Eli Lilly and Company, via DailyMed (US National Library of Medicine). HUMULIN 70/30 prescribing information, section 12 CLINICAL PHARMACOLOGY (12.1 Mechanism of Action and 12.2 Pharmacodynamics) (SPL set id e245e0c5-b2d6-418b-baa4-1c3324292885, effective 2026-06-17). 2026. Read the source
2. New England Journal of Medicine (Diabetes Control and Complications Trial Research Group). The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus.. 1993. PMID 8366922, DOI 10.1056/NEJM199309303291401. Read the source
3. The Lancet (UK Prospective Diabetes Study (UKPDS) Group). Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). UK Prospective Diabetes Study (UKPDS) Group.. 1998. PMID 9742976. Read the source
4. Diabetes Care. Comparison of basal insulin added to oral agents versus twice-daily premixed insulin as initial insulin therapy for type 2 diabetes.. 2005. PMID 15677775, DOI 10.2337/diacare.28.2.254. Read the source
5. Cochrane Database of Systematic Reviews. Short-acting insulin analogues versus regular human insulin for adult, non-pregnant persons with type 2 diabetes mellitus.. 2018. PMID 30556900, DOI 10.1002/14651858.CD013228. Read the source
6. Eli Lilly and Company, via DailyMed (US National Library of Medicine). HUMULIN 70/30 (insulin human) injectable suspension, section 11 DESCRIPTION (SPL set id e245e0c5-b2d6-418b-baa4-1c3324292885, effective 2026-06-17). 2026. Read the source
7. Eli Lilly and Company, via DailyMed (US National Library of Medicine). HUMULIN 70/30 prescribing information, section 7 DRUG INTERACTIONS, Table 1 in full (all four drug rows) (SPL set id e245e0c5-b2d6-418b-baa4-1c3324292885, effective 2026-06-17). 2026. Read the source
8. Eli Lilly and Company, via DailyMed (US National Library of Medicine). HUMULIN 70/30 prescribing information, section 5.3 Hypoglycemia, Risk Factors for Hypoglycemia (SPL set id e245e0c5-b2d6-418b-baa4-1c3324292885, effective 2026-06-17). 2026. Read the source
9. Eli Lilly and Company, via DailyMed (US National Library of Medicine). HUMULIN 70/30 prescribing information, section 7 DRUG INTERACTIONS, Table 1, drugs that may increase the risk of hypoglycemia (SPL set id e245e0c5-b2d6-418b-baa4-1c3324292885, effective 2026-06-17). 2026. Read the source
10. Eli Lilly and Company, via DailyMed (US National Library of Medicine). HUMULIN 70/30 prescribing information, section 5.7 Fluid Retention and Heart Failure with Concomitant Use of PPAR-gamma Agonists (SPL set id e245e0c5-b2d6-418b-baa4-1c3324292885, effective 2026-06-17). 2026. Read the source
11. Eli Lilly and Company, via DailyMed (US National Library of Medicine). HUMULIN 70/30 prescribing information, section 5.6 Hypokalemia (SPL set id e245e0c5-b2d6-418b-baa4-1c3324292885, effective 2026-06-17). 2026. Read the source
12. Eli Lilly and Company, via DailyMed (US National Library of Medicine). HUMULIN 70/30 prescribing information, section 2.2 Dosage Information (SPL set id e245e0c5-b2d6-418b-baa4-1c3324292885, effective 2026-06-17). 2026. Read the source
13. JAMA Internal Medicine. National estimates of insulin-related hypoglycemia and errors leading to emergency department visits and hospitalizations.. 2014. PMID 24615164, DOI 10.1001/jamainternmed.2014.136. Read the source
14. Eli Lilly and Company, via DailyMed (US National Library of Medicine). HUMULIN 70/30 prescribing information, section 6 ADVERSE REACTIONS body text (postmarketing reactions; Highlights excerpt excluded) (SPL set id e245e0c5-b2d6-418b-baa4-1c3324292885, effective 2026-06-17). 2026. Read the source
15. Eli Lilly and Company, via DailyMed (US National Library of Medicine). HUMULIN 70/30 prescribing information, section 10 OVERDOSAGE (SPL set id e245e0c5-b2d6-418b-baa4-1c3324292885, effective 2026-06-17). 2026. Read the source
16. The Lancet (UK Prospective Diabetes Study (UKPDS) Group). UKPDS 33 (FINDINGS section, hypoglycaemia and weight gain).. 1998. PMID 9742976. Read the source
17. New England Journal of Medicine (Action to Control Cardiovascular Risk in Diabetes Study Group). Effects of intensive glucose lowering in type 2 diabetes.. 2008. PMID 18539917, DOI 10.1056/NEJMoa0802743. Read the source
18. Cochrane Database of Systematic Reviews. (Ultra-)long-acting insulin analogues versus NPH insulin (human isophane insulin) for adults with type 2 diabetes mellitus.. 2020. PMID 33166419, DOI 10.1002/14651858.CD005613.pub4. Read the source
19. Cochrane Database of Systematic Reviews. (Ultra-)long-acting insulin analogues versus NPH insulin (human isophane insulin) for adults with type 2 diabetes mellitus (AUTHORS CONCLUSIONS).. 2020. PMID 33166419, DOI 10.1002/14651858.CD005613.pub4. Read the source
20. MannKind Corporation, via DailyMed (US National Library of Medicine). AFREZZA (insulin human) inhalation powder, BOXED WARNING (LOINC 34066-1) in Highlights (SPL set id 29f4637b-e204-425b-b89c-7238008d8c10, effective 2026-05-30). 2026. Read the source
21. Cochrane Database of Systematic Reviews. Short-acting insulin analogues versus regular human insulin for adult, non-pregnant persons with type 2 diabetes mellitus (AUTHORS CONCLUSIONS).. 2018. PMID 30556900, DOI 10.1002/14651858.CD013228. Read the source
22. New England Journal of Medicine (Diabetes Control and Complications Trial Research Group). The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus (METHODS section).. 1993. PMID 8366922, DOI 10.1056/NEJM199309303291401. Read the source
23. Diabetes Care. Comparison of basal insulin added to oral agents versus twice-daily premixed insulin as initial insulin therapy for type 2 diabetes (RESEARCH DESIGN AND METHODS).. 2005. PMID 15677775, DOI 10.2337/diacare.28.2.254. Read the source
24. Eli Lilly and Company, via DailyMed (US National Library of Medicine). HUMULIN 70/30 prescribing information, section 12.2 Pharmacodynamics (SPL set id e245e0c5-b2d6-418b-baa4-1c3324292885, effective 2026-06-17). 2026. Read the source