Clinical corner

Revisiting respiratory failure

by Richard D. Pinson, MD, FACP, CCS way pressure (BiPAP) nearly always means the patient has
The diagnosis and documentation of acute respiratory failure, but these measures are not required
respiratory failure continues to be challeng- for the diagnosis. Similarly, providing 40% or more supple-
ing for coders, documentation specialists, and mental oxygen implies that the physician is treating acute
physicians. Many physicians, including pulmon- respiratory failure since only a patient with acute respiratory
ologists, are unaware of the current clinical standards for diag- failure would need that much oxygen.
nosing acute respiratory failure and commonly overlook the Acute hypoxemic respiratory failure
presence of chronic respiratory failure. Yet they typically iden- The gold standard for the diagnosis of hypoxemic
tify multiple clinical criteria and provide appropriate manage- respiratory failure is an arterial pO2 on room air less than
ment for respiratory failure, which creates query opportunities. 60 mmHg measured by arterial blood gases (ABG). In the
In this article, we will discuss a variety of clinical indica- absence of an ABG, SpO2 measured by pulse oximetry on
tors for respiratory failure and identify a number of com- room air can serve as a substitute for the pO2: SpO2 of
mon documentation improvement opportunities. 91% equals pO2 of 60 mmHg. These criteria may not apply
Definition of acute respiratory failure to patients with chronic respiratory failure (e.g., severe chron-
Acute respiratory failure is classified as hypoxemic (low ic obstructive pulmonary disease [COPD]), because their
arterial oxygen levels), hypercapneic (elevated levels of car- room air pO2 is often less than 60 mmHg (SpO2 < 91%).
bon dioxide gas), or a combination of the two. In most cases Chronic respiratory failure patients are treated with
one or the other predominates. For ICD-9, these terms, supplemental oxygen on a continuous outpatient basis to keep
being “nonessential modifiers,” are irrelevant for code assign- arterial oxygen above these levels. However, if the baseline
ment. ICD-10, however, has codes that permit a distinction pO2 is known, a decrease by 10 mmHg or more indicates
(see Table 1), but the distinction is not a requirement and acute hypoxemic respiratory failure in such a patient.
queries for it will not alter its MCC classification. The clini- The P/F ratio
cal criteria for diagnosing acute respiratory ­failure are: The P/F ratio is a powerful objective tool to identify
»»Hypoxemic: Partial pressure of oxygen (pO2) level less acute hypoxemic respiratory failure at any time while the
than (<) 60 millimeter(s) of mercury (mmHg) (oxygen patient is receiving supplemental oxygen, a frequent problem
saturation [SpO2] < 91%) on room air, or pO2/frac- faced by documentation specialists where no room air ABG
tion of inspired oxygen (FIO2) (P/F) ratio < 300, or 10 is available, or pulse oximetry readings seem equivocal.
mmHg decrease in baseline pO2 (if known) The P/F ratio equals the arterial pO2 (“P”) from the
»»Hypercapnic: Partial pressure of carbon dioxide (pCO2) ABG divided by the FIO2 (“F”)—the fraction (percent)
>50 mmHg with pH < 7.35, or 10 mmHg increase in of inspired oxygen that the patient receives expressed as a
baseline pCO2 (if known) ­decimal (40% oxygen = FIO2 of 0.40). A P/F ratio less
than 300 indicates acute respiratory failure.
With the exception of the P/F ratio, these criteria have Most physicians have never heard of the P/F ratio, but
also been offered as assistance to coders and documentation it was validated and has been used in the context of acute
specialists for recognizing possible acute respiratory failure respiratory distress syndrome (ARDS) for many years, where
(see AHA’s Coding Clinic for ICD-9-CM, Third Quarter 1988, acute respiratory failure is called “acute lung injury.” A P/F
p. 7; and Second Quarter 1990, p. 20). ratio < 300 indicates mild ARDS, < 200 is consistent with
Management that requires endotracheal intubation and moderate ARDS, and < 100 is severe ARDS. The P/F ratio
mechanical ventilation or initiation of biphasic positive air- indicates what the pO2 would be on room air:
»»P/F ratio < 300 = a pO2 < 60 mm Hg on room air
For permission to reproduce part or all of this newsletter for external distribution or use in educational packets, please contact the Copyright Clearance Center at www.copyright.com or 978-750-8400.

20 January 2014 © 2014 HCPro, a division of BLR.

»»P/F ratio < 250 = a pO2 < 50 mm Hg on room air since only such a patient would need that much oxygen.
»»P/F ratio < 200 = a pO2 < 40 mm Hg on room air A nasal cannula provides oxygen at adjustable flow rates
in liters of oxygen per minute (L/min or LPM). The actual
As an example, suppose the pO2 is 90 mmHg on 40% FIO2 (percent oxygen) delivered by nasal cannula is some-
oxygen (FIO2 = .40). The P/F ratio = 90 divided by .40 what variable and less reliable than with a mask, but can be
= 225 (rather severe acute respiratory failure). The pO2 on estimated as shown in Table 3. The FIO2 derived from nasal
room air in this case would have been about 45 mmHg (well cannula flow rates can then be used to calculate the P/F ratio.
below the “cutoff ” of 60 mmHg). For example, a patient has a pO2 of 85 mmHg on ABG while
The validity of the P/F ratio is not limited to ARDS. receiving 5 L/min of oxygen. Since 5 L/min is equal to 40%
It simply expresses a consistent physiologic relationship oxygen (an FIO2 of 0.40), the P/F ratio = 85 divided by
between inspired oxygen and arterial pO2 regardless of 0.40 = 212.5 (clearly severe acute respiratory failure).
cause. Authoritative applications of the P/F ratio in set- Acute hypercapneic respiratory failure
tings other than ARDS include pneumonia and sepsis. The The hallmark of acute hypercapneic respiratory failure is
Infectious Disease Society of America and the American
Thoracic Society recognize a P/F ratio less than 250 as one
of the 10 criteria for “severe” community-acquired pneumonia
that may require admission to intensive care. The International Table 1: ICD-10-CM codes for
Sepsis Definition criteria (2001) and the Surviving Sepsis - respiratory failure
Severe Sepsis Guidelines (2008 and 2012) use a P/F ratio <
The following codes are applicable for respiratory failure
300 as an indicator of acute organ (respiratory) failure.
under ICD-10-CM:
SpO2 may be translated to pO2
The arterial pO2 measured by ABG is the definitive »» J96.0: Acute respiratory failure (MCCs)
ΩΩ J96.00: unspecified whether with hypoxia or
method for calculating the P/F ratio. However, when the
­hypercapnia
pO2 is unknown because an ABG is not available, the SpO2
ΩΩ J96.01: with hypoxia
measured by pulse oximetry can be used to approximate the
ΩΩ J96.02: with hypercapnia
pO2, as shown in Table 2. It is important to note that esti-
mating the pO2 from the SpO2 becomes unreliable when »» J96.1: Chronic respiratory failure (CCs)
ΩΩ J96.10: unspecified whether with hypoxia or
the SpO2 is greater than 97%.
­hypercapnia
For example, suppose a patient on 40% oxygen has a
ΩΩ J96.11: with hypoxia
pulse oximetry SpO2 of 95%. Referring to Table 2, SpO2
ΩΩ J96.12: with hypercapnia
of 95% is equal to a pO2 of 80 mmHg. The P/F ratio =
80 divided by 0.40 = 200 (quite severe acute respiratory »» J96.2: Acute and/on chronic respiratory failure (MCCs)
ΩΩ J96.20: unspecified whether with hypoxia or
failure). The patient may be stable receiving 40% oxygen, but
­hypercapnia
still has acute respiratory failure. If oxygen were withdrawn,
ΩΩ J96.21: with hypoxia
leaving her on room air, the pO2 would only be 42 mmHg
ΩΩ J96.22: with hypercapnia
(much less than 60 mmHg on room air).
Translating supplemental oxygen »» J96.9: Respiratory failure, unspecified
ΩΩ J96.90: Respiratory failure, unspecified, (unspecified
Supplemental oxygen may be administered by mask
whether with hypoxia or hypercapnia)
or nasal cannula. A Venturi mask (Venti-mask) delivers a
ΩΩ J96.91: Respiratory failure, unspecified with hypoxia
controlled flow of oxygen at a specific fixed concentration
ΩΩ J96.92: Respiratory failure, unspecified with hypercapnia
(FIO2): 24%, 28%, 31%, 35%, 40%, and 50%. The non-
(excludes newborn, postprocedural, ARDS, respiratory
rebreather (NRB) mask is designed to deliver approximately
arrest, and cardiorespiratory failure)
100% oxygen. Providing 40% or more ­supplemental oxygen
implies that the physician is treating acute respiratory failure
For permission to reproduce part or all of this newsletter for external distribution or use in educational packets, please contact the Copyright Clearance Center at www.copyright.com or 978-750-8400.

© 2014 HCPro, a division of BLR. January 2014 21

elevated pCO2 due to retention/accumulation of carbon diox- classified as acute-on-chronic respiratory failure if properly
ide gas resulting in an acidic pH less than 7.35. There are many documented.
causes, but severe COPD is the most common. Physicians can Chronic respiratory failure
establish a diagnosis by viewing a pCO2 greater than 50 mmHg Chronic respiratory failure is very common in patients
with a pH less than 7.35. If the pH is greater than 7.35, the with severe COPD and other chronic lung diseases such as
patient has chronic (not acute) respiratory failure. cystic fibrosis and pulmonary fibrosis. It is characterized
Physicians often identify this clinical condition as “respira- by a combination of hypoxemia, elevated pCO2, elevated
tory acidosis,” which is the same thing as acute hypercapneic bicarbonate level, and normal pH (7.35–7.45). The most
respiratory failure. Unfortunately, the code for “respiratory aci- important tip-off to chronic respiratory failure is chronic
dosis” is 276.2, which is a CC, in contrast to the MCC status of dependence on supplemental oxygen (“home O2”).
acute respiratory failure—hence the need for clarification. Patients who qualify for home O2 almost always have
Also, if the baseline pCO2 is known, an increase of 10 chronic respiratory failure. Another clue is an elevated bicar-
mmHg or more indicates acute hypercapneic respiratory bonate level on the basic metabolic panel (BMP) in a COPD
failure. Finally, an exacerbation of symptoms requiring an patient, especially helpful when no ABG was obtained.
increase in chronic supplemental oxygen indicates an “acute For example, consider a patient admitted with CHF
exacerbation” of chronic respiratory failure, which would be exacerbation and a history of severe COPD. ABG on room
air shows pH 7.40, pCO2 52 mmHg, and pO2 70 mmHg;
bicarbonate level on BMP is elevated at 42. This is classic
chronic respiratory failure: normal pH, elevated pCO2 and
Table 2: Conversion of SpO2 to pO2 bicarbonate, with hypoxemia—but no acute criteria.
The following chart illustrates the conversion of SpO2 Acute-on-chronic respiratory failure
to pO2: When a patient experiences an acute exacerbation or
decompensation of chronic respiratory failure, he has
SpO2 (percent) pO2 (mmHg)
­“acute-on-chronic” respiratory failure. It is recognized by any
of the following:
85 50
»»Worsening symptoms
86 51
»»Greater hypoxemia (hypoxemic)
87 52 »»Elevated pCO2 with pH < 7.35 (hypercapneic)
88 54 During an acute exacerbation, acidic carbon dioxide
89 56 (pCO2) may accumulate rapidly (“CO2 retention”), causing
90 58 acidosis with a pH < 7.35 (acute hypercapneic respiratory
failure). This would be acute-on-chronic respiratory failure.
91 60
Worsening of symptoms requiring an increase in supplemen-
92 64
tal oxygen also indicates an “acute exacerbation” of chronic
93 68 respiratory failure.
94 73 Use hypoxemic criteria (pO2, SpO2, and P/F ratio) in
95 80 patients with chronic respiratory failure with caution. Many
96 90 of these patients always have a pO2 < 60 mmHg on room
97 110
air, which is the reason they use supplemental oxygen. For
such patients, the pO2/SpO2 criterion can be applied, not
Note: Estimating the pO2 from the SpO2 becomes unreli-
on room air, but while receiving their usual supplemental
able when the SpO2 is greater than 97%.
oxygen flow. Why? Because home O2 is adjusted to maintain
a pO2 > 60 mmHg (SpO2 > 91%). Therefore, if the pO2
is < 60 mmHg on the usual supplemental oxygen flow rate,
For permission to reproduce part or all of this newsletter for external distribution or use in educational packets, please contact the Copyright Clearance Center at www.copyright.com or 978-750-8400.

22 January 2014 © 2014 HCPro, a division of BLR.

acutely decompensated respiratory failure has occurred. of a code for post-procedural respiratory failure. For exam-
Do not use the P/F ratio to diagnose acute-on-chronic ple, something like: “acute respiratory failure in the postop
respiratory failure since it is typically < 300 in these patients setting primarily due to preexisting CHF.”
at baseline. It may be used to monitor the patient’s clinical Summary
progress over time; if it keeps dropping, the patient is getting Understanding the pathophysiology and authoritative
worse and needs more aggressive treatment. clinical criteria for the several types of respiratory failure
Post-procedural respiratory failure empowers coders and documentation specialists to con-
The diagnosis of respiratory failure following surgery has fidently recognize, query, validate, and compliantly code
profound regulatory and quality of care implications. If iden- these conditions. The two basic types of respiratory failure
tified as “postop,” “due to,” or “complicating” a procedure, are hypoxemic and hypercapneic, sometimes occurring in
respiratory failure is classified as one of the most severe, life- combination. The distinction is clinically important but not
threatening, reportable surgical complications a patient can have. required for correct coding using either ICD-9 or ICD-10.
This diagnosis adversely affects quality scores for both the The P/F ratio is a powerful diagnostic, prognostic, and clin-
hospital and the surgeon. On the other hand, the diagnosis and ical management tool: P/F ratio < 300 indicates acute respira-
coding of post-procedural respiratory failure (an MCC) often tory failure. However, the acute hypoxemic criteria (pO2/SpO2
results in large payment increases for hospitals. If improperly and P/F ratio) must be applied with caution to the diagnosis
diagnosed without firm clinical grounds, it may become the of acute-on-chronic respiratory failure since they are frequently
basis for regulatory or contractual audits, penalties, sanctions, abnormal in the patient’s stable, chronic, baseline state.
and even legal action affecting the hospital and the physician. Carefully consider the implications of diagnosing and
Post-procedural respiratory failure is a lucrative Recovery coding post-procedural respiratory failure; clarify any poten-
Auditor target. Facilities should have a policy that governs tial relationship to preexisting conditions when present.
the coding of any condition (including respiratory failure)
not supported by clinical criteria in the medical record.
Editor’s note
To validate the diagnosis, the patient must have acute Pinson is a certified coding specialist and a principal partner at HCQ Consulting
pulmonary dysfunction requiring nonroutine aggressive (www.hcqconsulting.com). He is coauthor of The CDI Pocket Guide and the CDI+ and
CDI+MD mobile apps.
measures. A patient who requires a short period of ventilator
support during surgical recovery does not have acute respira-
tory failure; do not assign a code in this instance. The same
is true for any duration of mechanical ventilation that is
usual or expected following the type of surgery performed, Table 3: Conversion of nasal cannula
unless there truly is underlying acute pulmonary dysfunction. oxygen flow rate to FIO2
A further difficulty arises because coding rules inexplicably
call for coding of postop respiratory failure as a complication The following figures illustrated the conversion of nasal
of care even when terms that seem clinically innocuous to phy- cannula oxygen flow rate to FIO2:
sicians are used in the postop setting, such as pulmonary insuf-
ficiency (acute or not) and acute respiratory insufficiency. To Flow Rate FIO2
avoid confusion and improper code assignment, instruct your 1 L/min 24%
physicians not to use such terms in the postoperative setting 2 L/min 28%
unless the patient actually has acute respiratory failure. 3 L/min 32%
If the patient has acute respiratory failure following sur- 4 L/min 36%
gery, but it is truly due to, primarily the result of, or related 5 L/min 40%
to a preexisting medical condition (such as COPD, CHF, 6 L/min 44%
a neuromuscular disorder, etc.), ask the physician to clearly
document this connection to avoid the incorrect assignment
For permission to reproduce part or all of this newsletter for external distribution or use in educational packets, please contact the Copyright Clearance Center at www.copyright.com or 978-750-8400.

© 2014 HCPro, a division of BLR. January 2014 23

 December ACDIS ICD-10-CM/PCS query preparation survey results

1. To date, have your CDI staff received information to raise their awareness of ICD-10 implementation and documentation
improvement needs?
Answer options Response percent Response count
Yes 87.7% 100
No 12.3% 14
Other (please specify) 0
answered question 114
skipped question 0

2. To date, have your CDI staff received ICD-10 training on the code set?
Answer options Response percent Response count
Yes 67.5% 77
No 32.5% 37
Other (please specify) 0
answered question 114
skipped question 0

3. To date, have your CDI staff assisted with the ICD-10 education of physicians?
Answer options Response percent Response count
Yes 31.5% 35
No 68.5% 76
Other (please specify) 3
answered question 111
skipped question 3

4. Have you to date, or do you plan to, train CDI staff on the actual ICD-10 code set?
Answer options Response percent Response count
Yes 90.8% 99
No 9.2% 10
Other (please specify) 7
answered question 109
skipped question 5

For permission to reproduce part or all of this newsletter for external distribution or use in educational packets, please contact the Copyright Clearance Center at www.copyright.com or 978-750-8400.

24 January 2014 © 2014 HCPro, a division of BLR.

 5. Which of the following query templates do you use in your organization today?
Answer options Response percent Response count
Anemia 89.5% 85
Angina 36.8% 35
CAD 32.6% 31
Cause and effect 54.7% 52
Coma 21.1% 20
Complication 47.4% 45
Diabetes 50.5% 48
Diabetes, controlled or uncontrolled 41.1% 39
Fracture 31.6% 30
Heart failure 95.8% 91
Liver failure 14.7% 14
Malnutrition 88.4% 84
Renal failure 84.2% 80
Respiratory failure 82.1% 78
Sepsis 90.5% 86
Other (please specify) 26
answered question 95
skipped question 19

6. Have you conducted an inventory of your physician queries by type and frequency?
Answer options Response percent Response count
Yes 37.5% 42
Yes, by type 11.6% 13
Yes, by frequency 8.9% 10
No 34.8% 39
Don’t know 7.1% 8
Other (please specify) 1
answered question 112
skipped question 2

For permission to reproduce part or all of this newsletter for external distribution or use in educational packets, please contact the Copyright Clearance Center at www.copyright.com or 978-750-8400.

© 2014 HCPro, a division of BLR. January 2014 25

 December ACDIS ICD-10-CM/PCS query preparation survey results (cont.)

7. Have you started to audit (review) and update queries for ICD-10 language changes?
Answer options Response percent Response count
Yes 18.2% 20
Yes, we have audited our queries 8.2% 9
Yes, we have audited our queries and updated them for ICD-10 10.9% 12
No 30.9% 34
No, but we plan do this in the first quarter of 2014 29.1% 32
Don’t know 2.7% 3
Other (please specify) 7
answered question 110
skipped question 4

8. Does your compliance department review new/updated physician queries to ensure they are compliant?
Answer options Response percent Response count
Yes 29.5% 33
No 53.6% 60
Don’t know 17% 19
Other (please specify) 5
answered question 112
skipped question 2

9. Do your physicians review new/updated queries?

Answer options Response percent Response count
Yes 5.5% 6
Yes, our physician advisor reviews all new/updated queries 20% 22
Yes, our physicians review any new/updated queries by specialty 5.5% 6
No 61.8% 68
Don’t know 7.3% 8
Other (please specify) 7
answered question 110
skipped question 4

For permission to reproduce part or all of this newsletter for external distribution or use in educational packets, please contact the Copyright Clearance Center at www.copyright.com or 978-750-8400.

26 January 2014 © 2014 HCPro, a division of BLR.

 10. Does your CDI program staff meet regularly with your HIM/coding staff?
Answer options Response percent Response count
Yes 19.1% 21
Yes, weekly 10% 11
Yes, monthly 30% 33
Yes, quarterly 15.5% 17
No 24.5% 27
Don’t know 0.9% 1
Other (please specify) 5
answered question 110
skipped question 4

For permission to reproduce part or all of this newsletter for external distribution or use in educational packets, please contact the Copyright Clearance Center at www.copyright.com or 978-750-8400.

© 2014 HCPro, a division of BLR. January 2014 27