What Is the ICD-10 Code for Peripheral Artery Disease (PAD)?

The primary ICD-10-CM code for peripheral artery disease is I70.2, which covers atherosclerosis of the native arteries of the extremities. In practice, though, no one submits a bare I70.2. The U.S. clinical modification of ICD-10 demands subcodes that specify which leg is affected, what symptoms are present, and how severe the disease has become. The full code range spans I70.2 through I70.7, with the higher numbers reserved for disease in bypass grafts rather than native vessels. Getting this right matters for reimbursement, clinical tracking, and research, yet studies consistently find that coding inconsistencies lead to real problems in how PAD is identified and studied.

How the Code Structure Works

ICD-10-CM breaks PAD into a layered system. The base code I70.2 means atherosclerosis of the native (original, non-grafted) arteries of the extremities. From there, additional digits capture clinical detail. The fourth character indicates the type of vessel involved, while subsequent characters specify laterality and clinical presentation. The subcodes capture four key clinical stages: intermittent claudication (pain during walking that resolves with rest), rest pain (pain even without exertion), ulceration, and gangrene. Countries outside the United States have taken their own approaches to this. Germany’s ICD-10-GM, for instance, maps the I70.2 subcodes directly to the Fontaine classification stages, and researchers have used this granularity to build amputation risk models. The U.S. system captures similar clinical detail but organizes it differently.

A common example helps illustrate the specificity involved. A patient with intermittent claudication in the right leg would receive code I70.211. If that same patient progressed to rest pain, the code would shift to I70.221. Ulceration in the left leg with a breakdown at the calf level would carry a different code still. Each clinical scenario maps to its own unique entry, meaning the coding system can theoretically paint a fairly detailed picture of a patient’s disease status at any point in time.

Beyond Native Arteries

The I70.2 family covers disease in native arteries, but many PAD patients have had prior surgical interventions. When atherosclerosis develops in a bypass graft rather than the original artery, the coding shifts to higher categories. I70.3 covers atherosclerosis of unspecified bypass grafts, I70.4 covers autologous vein grafts (grafts made from the patient’s own vein), I70.5 covers nonautologous biological grafts, I70.6 covers nonbiological (synthetic) grafts, and I70.7 covers other types. Each of these carries the same downstream specificity requirements for laterality, symptoms, and severity. This distinction is clinically meaningful because disease in a bypass graft often indicates progression and may require different treatment strategies than disease in a native vessel.

Secondary interventions are relatively common after both endovascular and open revascularization procedures. One large study found that patients who initially received an endovascular procedure underwent subsequent interventions about 7 to 9 percent more often than those who had open surgery, a gap that persisted over four years of follow-up.² Each of those subsequent procedures generates its own set of diagnosis and procedure codes, so a single patient’s coding record can become complex over time.

How Well Do the Codes Actually Work

Assigning the right code is one thing. Whether the codes reliably capture what is happening clinically is another question entirely. One of the most important distinctions in PAD coding is between intermittent claudication, the milder form where walking triggers leg pain, and chronic limb-threatening ischemia (CLTI), the severe end of the spectrum where blood flow is so compromised that tissue loss and amputation become real risks.³ A study of over 54,000 patients who underwent peripheral vascular interventions found that a predefined set of ICD-10 codes could distinguish claudication from CLTI with about 81 percent sensitivity and 82 percent specificity. Machine-learning models pushed sensitivity to nearly 89 percent and overall agreement to about 84 percent.⁴

Those numbers are encouraging but far from perfect. Roughly one in five patients gets classified into the wrong severity group based on codes alone. That error rate has real downstream consequences: treatment decisions, risk stratification, and quality benchmarks all depend on accurate severity classification. And the problem compounds in research settings, where administrative codes are often the only data available for large population studies. A scoping review of ninety observational studies on PAD found that 40 percent did not even specify the grade of PAD being studied, and among those that did, ICD code usage was remarkably inconsistent and overlapping.⁵

Why Risk Factors Complicate the Coding Picture

PAD is fundamentally an atherosclerotic disease, meaning it shares the same root cause as coronary artery disease and cerebrovascular disease: the buildup of plaque inside artery walls.⁶ It affects an estimated 200 million people worldwide.⁷ The major modifiable risk factors are smoking, high blood pressure, high cholesterol, and diabetes, and their effects stack. A large cohort study found that each additional risk factor roughly doubled a person’s PAD risk. Someone with all four risk factors had an incidence rate more than twenty times higher than someone with none.⁸

This matters for coding because PAD rarely exists in isolation. A patient coded with I70.211 for claudication in the right leg is very likely to also carry codes for hypertension, diabetes, dyslipidemia, and possibly coronary artery disease. PAD itself independently raises the risk of stroke, heart failure, and cardiovascular death. One large nationwide study found that PAD roughly doubled the risk of cardiovascular death, with similar independent elevations in stroke and heart failure hospitalization risk.⁹ The coding record for a PAD patient, then, is never just about the legs. It reflects a systemic vascular disease with consequences across the entire cardiovascular system.¹⁰

The Asymptomatic Problem

One of the trickiest coding scenarios involves patients who have PAD but do not know it. The disease can narrow arteries significantly before producing any noticeable symptoms, particularly in people who are sedentary enough that they never push their legs hard enough to trigger claudication. The ankle-brachial index (ABI), a simple comparison of blood pressure at the ankle versus the arm, is the standard screening tool. A systematic review found that ABI detects significant arterial narrowing with about 61 percent sensitivity and 92 percent specificity.¹¹

Despite this, the U.S. Preventive Services Task Force concluded that there is not enough evidence to recommend routine ABI screening in asymptomatic adults. The task force found that treatment benefits for screen-detected PAD are not established, that adding the ABI to standard cardiovascular risk scores probably has limited predictive value, and that one large trial showed low-dose aspirin did not prevent cardiovascular events in people with screen-detected low ABI and may have increased bleeding risk.¹²¹³ The practical consequence for coding: many people with genuine PAD never receive a diagnosis or an ICD-10 code at all, which means prevalence estimates drawn from claims data almost certainly undercount the true burden of the disease.

Sex Differences in Presentation and Diagnosis

The underdiagnosis problem is especially pronounced in women. PAD has traditionally been viewed as a disease of older men who smoke, but the reality is more complicated. Women with PAD are less likely to present with the classic symptom of intermittent claudication. They are more likely to be asymptomatic, to have atypical exercise-related symptoms, or to have overlapping musculoskeletal complaints that mask the underlying vascular problem. A systematic review encompassing nearly two million PAD patients confirmed this pattern: women were significantly less likely than men to present with claudication but were more prone to developing critical limb ischemia, with some studies suggesting a two-fold higher prevalence of that severe presentation compared to men.¹⁴

This has coding implications beyond the obvious clinical ones. If women are underdiagnosed until the disease reaches an advanced stage, their initial ICD-10 codes are more likely to reflect rest pain, ulceration, or gangrene rather than early claudication. That skews administrative data in ways that make it look like women have worse disease at baseline, when in reality the disease has simply gone unrecognized longer. Researchers who rely on ICD-10 codes to study PAD epidemiology need to account for this systematic bias.

When Treatment Codes Enter the Picture

PAD management generates its own coding ecosystem beyond the diagnosis codes. Guideline-directed medical therapy, which typically includes antiplatelet agents, statins, antihypertensives, and blood sugar control for diabetic patients, carries its own documentation requirements. The evidence for sticking with these medications is strong: patients who did not receive guideline-directed therapy after a peripheral vascular intervention had a 37 percent higher risk of death and a 20 percent higher risk of amputation compared to those who did.¹⁵

When medical therapy is insufficient and revascularization becomes necessary, the coding branches further. Procedure codes differentiate between endovascular interventions (stent placement, balloon angioplasty) and open surgical bypass, and the specific bypass graft material matters both clinically and for code selection. In diabetic patients specifically, the choice between stent revascularization and bypass surgery has been studied extensively, with large database analyses tracking outcomes across thousands of patients.¹⁶ Each treatment pathway leaves its own trail in the coding record, and subsequent disease in a bypass graft shifts the diagnosis from the I70.2 family to I70.3 through I70.7, as noted earlier.

Social Factors and the Geography of Amputation

Where you live and what resources you have access to shape PAD outcomes in ways that coding data can reveal but not fix. Sociodemographic disadvantages including low income, limited education, lack of insurance, and rural geography are all associated with higher rates of PAD development and, critically, higher rates of lower extremity amputation.¹⁷ Race and ethnicity compound these disparities. The same ICD-10 code for critical limb ischemia can lead to dramatically different outcomes depending on whether the patient has access to a vascular specialist, lives near a hospital with endovascular capabilities, or has insurance that covers the full continuum of care.

Amputation codes themselves carry weight in quality measurement. Hospitals and health systems increasingly track major and minor amputation rates as performance indicators. A PAD patient’s coding journey from I70.211 (claudication) to I70.261 (gangrene) to a procedure code for below-knee amputation represents a cascade that, in many cases, could have been interrupted with earlier intervention. Researchers studying these patterns rely on accurate ICD-10 coding to identify where the system breaks down, which is why the coding inconsistencies documented in the scoping review mentioned earlier are so concerning for health equity research.

Diagnostic Overlap and Coding Pitfalls

PAD does not always present in a textbook fashion, and the ICD-10 system has some blind spots worth knowing about. Leg pain during walking can stem from spinal stenosis, venous insufficiency, musculoskeletal problems, or neuropathy rather than arterial disease. A clinician who codes leg pain as claudication without confirming arterial involvement creates a false positive in the administrative record. Conversely, a patient with genuine PAD whose symptoms are attributed to a musculoskeletal cause becomes a false negative.

The validation study of over 54,000 patients undergoing vascular interventions found roughly 81 percent total agreement between ICD-10 codes and gold-standard registry definitions.⁴ That sounds respectable until you consider that these were patients already undergoing a vascular procedure, meaning the diagnosis was not in doubt. In the general clinical population, where PAD has to be distinguished from a wider field of possible diagnoses, accuracy is likely lower. The England-based study comparing diagnostic codes across administrative and clinical databases found similar concerns and noted that the base ICD-10 system (used in most countries outside the United States) lacks the granular subcodes that the U.S. and German modifications provide.¹

If you are a patient reviewing your own medical records or explanation of benefits, the key thing to look for is whether the code matches your actual symptoms and which leg is affected. An incorrect laterality code or a severity level that does not reflect your current clinical status can cause problems with insurance authorization for treatments, particularly for advanced interventions that require documentation of disease severity. If the code does not match your experience, it is worth raising with your provider’s billing office before it becomes entrenched in your record.