13.2 Sepsis Definitions, SEP-1, and Coding

Sepsis: clinical assessment, quality abstraction, and coding

Sepsis review involves several related decisions with different purposes. Clinicians assess and treat an evolving illness. A quality abstractor applies a particular measure's definitions and timing rules. A coder applies ICD-10-CM to the established documentation. A payer may separately review clinical support under applicable coverage and claim-review rules. One decision cannot automatically substitute for the others.

For CDIP questions, first identify which decision is being tested. A screening alert does not establish a diagnosis; a quality-measure threshold is not a universal diagnostic rule; and a documented sepsis diagnosis does not by itself describe every required code or relationship.

Understand the clinical frameworks

Earlier consensus definitions used infection with systemic inflammatory response syndrome (SIRS), with “severe sepsis” describing associated organ dysfunction and septic shock describing persistent sepsis-induced hypotension despite adequate resuscitation. The original four-item SIRS framework used at least two of: temperature above 38°C or below 36°C; heart rate above 90/min; respiratory rate above 20/min or arterial carbon dioxide below 32 mm Hg; and white blood cells above 12,000/mm³, below 4,000/mm³, or more than 10% immature bands. Do not substitute a later measure's temperature threshold and label it the original definition.

The 2016 Sepsis-3 consensus instead emphasizes a dysregulated response to infection causing life-threatening organ dysfunction. Its clinical operationalization uses an acute increase of at least two points in the original SOFA score attributable to infection. Known preexisting organ dysfunction matters when determining the baseline. SOFA examines respiratory, coagulation, liver, cardiovascular, neurologic, and renal function; a raw score without context does not establish which findings are acute or infection-related.

Under that consensus, septic shock combines sepsis, a persistent vasopressor requirement to maintain MAP at least 65 mm Hg, and lactate above 2 mmol/L despite adequate volume resuscitation. The definition does not specify that every patient must first receive an identical 30 mL/kg volume.

Sepsis-3 regarded “severe sepsis” as redundant clinical terminology. ICD-10-CM still retains severe sepsis coding. Therefore, do not delete R65.2- from coding practice because a clinical paper changed its terminology, or automatically translate every Sepsis-3 assessment into a complete severe-sepsis code set without evaluating documentation and relationships.

Screening and current treatment guidance

The three qSOFA findings—respiratory rate at least 22/min, systolic pressure at most 100 mm Hg, and altered mentation—were described as a prompt to recognize adults with suspected infection at greater risk of poor outcomes. They are not a stand-alone diagnostic test. A score below two does not exclude sepsis, and a score of two does not prove it.

The final 2026 Surviving Sepsis Campaign adult guidance recommends NEWS, NEWS2, MEWS, or SIRS over qSOFA when using a single hospital screening tool. It also states that sepsis cannot be established or excluded with one biomarker or diagnostic test. Negative cultures, a normal lactate, or absence of one inflammatory sign must be interpreted in the entire clinical course. Cultures may be uninformative; competing noninfectious explanations must also be considered.

The 2026 guidance addresses urgent evaluation and treatment, antimicrobial stewardship, hemodynamic support, and reassessment. Its initial fluid recommendation for sepsis-induced hypoperfusion or septic shock is conditional, with patient context and ongoing reassessment explicitly relevant. A treatment recommendation, its evidence strength, and a measure abstraction rule are distinct. CDI should preserve the clinician's rationale and actual treatment record rather than request a diagnosis to justify a protocol retrospectively.

SEP-1 uses versioned abstraction rules

CMS QualityNet lists manual version 5.18a for discharges in calendar 2026 and 5.19 for calendar 2027. These calendar-year periods differ from the fiscal-year ICD-10-CM code cycle. Select each source by its own applicability date.

Both manuals' SEP-1 numerator statements describe the following structure:

This is a guide to the numerator's structure, not the complete abstraction algorithm. Presentation times, qualifying evidence, acceptable timing windows, medications, exclusions, and data-element instructions must also be applied. For example, version 5.18a has different SIRS thresholds for specified pregnancy/post-delivery circumstances and nonpregnant patients. Its nonpregnant temperature threshold is above 38.3°C or below 36°C, rather than the original SIRS threshold above 38°C.

The fluid data element in 5.18a also allows a lesser ordered volume when its documentation requirements are satisfied, including the specified volume and reason in the same source and the other applicable conditions. An order below 30 mL/kg is therefore not an automatic measure failure. Conversely, a vague reference to a fluid restriction does not establish compliance with every requirement. Use the applicable manual and actual documentation.

Suppose severe sepsis presentation is abstracted at noon and initial hypotension at 15:00. The numerator's fluid interval follows the applicable hypotension or shock trigger; it is not automatically a noon-to-15:00 interval merely because the first three components use severe sepsis presentation. This distinction is about measure abstraction and does not authorize delaying urgent care until an abstraction deadline.

Apply FY 2027 sepsis coding separately

Under I.C.1.d, identify the systemic infection, any localized infection, associated acute organ dysfunction, shock, and special circumstances.

• Sepsis with a localized infection: when the reason for admission is sepsis or severe sepsis and the localized infection, sequence the systemic infection first and the localized infection second. If admission was for a localized infection and sepsis developed afterward, the localized infection is sequenced first. POA alone does not override the principal-diagnosis definition or a special sequencing rule.
• Severe sepsis: code the underlying systemic infection and R65.2-, with additional codes for associated acute organ dysfunction. If the organism is not documented, A41.9 may identify the systemic infection. R65.2- cannot be principal.
• Unclear organ-dysfunction relationship: sepsis plus an unrelated acute condition does not justify severe-sepsis coding. If the relationship is unclear, query. Do not infer the relationship from a score or from simultaneous occurrence.
• Septic shock: follow the systemic-infection and shock sequencing rules and code other acute organ dysfunction as applicable. Postprocedural septic shock follows its special instructions, including T81.12- and not an additional R65.21.

Special circumstances can change the usual sequence. Postprocedural surgical-site infections and infection-associated hemolytic-uremic syndrome have specific instructions. If infection-associated HUS is the reason for admission, I.C.1.d.9 directs D59.31 first. Obstetric and newborn cases require their respective chapter guidance. A single “sepsis always first” flowchart is unsafe.

Two frequent clarification problems

Urosepsis: I.C.1.d.1 explicitly states that this nonspecific term is not synonymous with sepsis and has no default code in the Alphabetic Index. Query for clarification. Do not default it to N39.0, A41.9, or severe sepsis. Separately established UTI documentation can support its own coding; the word “urosepsis” supplies no default diagnosis.

Negative or inconclusive blood cultures: the official guideline states that these do not preclude sepsis in a patient with clinical evidence and directs provider clarification. Neither a positive culture nor a negative culture independently settles the entire sepsis assessment. Do not use negative results as an automatic deletion rule.

Query the unresolved clinical issue

Build a dated timeline of suspected infection, baseline and changing organ function, relevant findings, actual treatment, and the provider's interpretation. Distinguish “creatinine increased” from documented AKI and “lactate elevated” from documented lactic acidosis. Ask the appropriate involved provider about the diagnosis or relationship that remains unclear.

For instance, the provider documents sepsis and AKI but also states that the AKI resulted from an unrelated obstruction. That record does not support linking the AKI to sepsis merely to assign R65.20. If other documentation creates uncertainty about the relationship, request clarification using both supporting and contrary evidence. If the unrelated cause is clear and consistent, apply the existing record.

Do not promise that a particular phrase, SOFA value, or duration of organ dysfunction guarantees payment or defeats a denial. Review actual payer reasoning and applicable sources, preserve the provider's clinical judgment, and keep query content free of reimbursement pressure. Accuracy may support confirmation, a different diagnosis, or removal of an unsupported diagnosis through proper clarification.

Sources: Sepsis-3 consensus; final 2026 SSC adult guidance; CMS QualityNet manuals, 5.18a and 5.19, SEP-1 and relevant data elements; FY 2027 ICD-10-CM I.C.1.d; 2026 compliant query practice brief.