Cynthia Rothermel Pred Apnea Hypopnea Index from Overnight Pulse

Document Sample
Cynthia Rothermel Pred Apnea Hypopnea Index from Overnight Pulse Powered By Docstoc
					Prediction of the Apnea-
Hypopnea Index form
Overnight Pulse Oximetry.

               Magalang, UJ. Dmochowski, J. Veeramachaneni, S.
                       Draw, A. Mador MJ. El-Sohl, A. Grant, BJB.

Presentation by: Cynthia Rothermel
• Article can be found in
  CHEST 124(5): 1694-701
  2003 Nov.
• A Peer Reviewed Journal
• Copies of this article will be
  distributed to all class
 Understand the Apnea-Hypopnea
  Index (AHI)
 Understand how overnight pulse
  oximetry is used to determine
  several indexes used in the
  prediction of the AHI.
 Understand the relationship of
  these indexes to the AHI
 Decide if pulse oximetry can be
  used to predict the AHI without
Sleep Apnea

• Affects between 2 and 4% of
  middle-aged population
• Polysomnography (sleep
  study) is considered the “Gold
  Standard” for diagnostic
• Polysomnography is costly,
  time consuming, and difficult
  to repeat.
The derived indexes from
overnight pulse oximetry
• Δ (Delta) index—a measure of
  variability of the SaO2.
• # of oxyhemoglobin
  desaturations below 2%, 3%
  and 4% of baseline per hour
• Cumulative time spent below
  90%, 88%, 86%, 84%, 82%, and
  80% saturation proportional to
  time recorded.
Apnea-Hypopnea Index

• Defined as the number of
  apneas added to the number
  of hypopneas per hour of
• Standard criteria is considered
  > 15/hour.
This study
• Examines whether these indexes
  are a good indicator of
  Obstructive Sleep Apnea
• Examines the correlation between
  these indexes and the Apnea-
  Hypopnea Index
• Uses 2 independent sleep labs
• Looks at the results from 516
  patients undergoing
Study group eligibility
• Pts that underwent overnight
• Over 18 years of age
• No supplemental oxygen used
  during the sleep study
• No same night CPAP titration
• Informed consent was required of
  pts at Associated Sleep Center, but
  not from those at Buffalo Veterans
  Affairs Medical Center
Data collection
• Pulse oximetry was collected
  during polysomnography
  testing with a probe placed
  on the pt’s finger
• Apnea defines as lack of air
  flow for 10 sec or longer
• Hypopnea defined as visible
  reduction in air flow for at least
  10 sec with a 4% desaturation
  or an EEG arousal
Desaturation Events

• Defined as a decrease of the
  set amount from baseline for
  at least 10 secs at a rate of >
• The SaO2 return to within 1% of
  baseline within 60 secs or
  increase from lowest point by
  1.5 times or more.
Delta (Δ) Index

• Calculated as the average of
  absolute differences of the
  SaO2 between successive 12
  sec. intervals.
• Quantifies oxygen saturation
Data collection
• Sleep studies scored by a
  single person, blinded to the
  study, in each sleep lab.
• Divided into 3 groups:
     derivation group – 224
     validation group 1 – 101
     validation group 2 – 191
• Adjusted logarithmically for
Corrections for Error
• Concern that the results would
  depend on the cases selected
• Used multivariate regression
  techniques to develop
  prediction model
• Made sure the various indexes
  correlated with each other
• Prediction model validated in
  2 independent facilities
Accuracy of the derived indexes from pulse
             oximetry values
Oximetry Index                  c Index      95% CI

Δ{Delta} index                  0.881     0.8384–0.9236
Desaturation index (3% level)   0.873     0.8284–0.9176
Desaturation index (2% level)   0.868     0.8222–0.9138
Desaturation index (4% level)   0.852     0.8010–0.9030
Cumulative time below 90%       0.772     0.7096–0.8344
Cumulative time below 88%       0.757     0.6924–0.8216
Cumulative time below 86%       0.723     0.6566–0.7894
Cumulative time below 84%       0.692     0.6262–0.7578
Cumulative time below 82%       0.674     0.6108–0.7372
Cumulative time below 80%       0.647     0.5854–0.7086
 The positive likelihood
 ratios in the derivation
group (top, A), validation
 group 1 (center, B), and
    validation group 2
(bottom, C) according to
  the severity of the AHI
derived from multivariate
prediction models using
 the {Delta} index alone
 (filled triangles) vs the
aggregated model (filled
Model of predicted from P.O. to actual
AHI values post study for comparison
What does it all mean?
• The Δ index is the a good
  indicator of OSA.
• The desaturation events are
  closely correlating to the Δ
• All result taken together are
  the best indicator of the AHI
• The Δ index is sensitive, yet not
  specific as an individual result
Overnight pulse oximetry

How do the indexes derived
   from pulse oximetry compare
   to the AHI found during
Is this a reasonable alternative
   to polysomnography testing?
• Effect of body mass index on overnight oximetry
  for the diagnosis of sleep apnea.
  Nakano H. Ikeda T. Hayashi M. Ohshima E. Itoh M.
  Nishikata N. Shinohara T.
  Respiratory Medicine. 98(5):421-7, 2004 May.
• Sleep on the cheap: the role of overnight
  oximetry in the diagnosis of sleep apnoea
  hypopnoea syndrome. Bennett, JA; Kinnear, WJM.
  Thorax 54:95/-959 1999 Nov.
• Prediction of the Apnea-Hypopnea Index From
  Overnight Pulse Oximetry. Magalang, UJ;
  Dmochowski, J; Veeramachaneni, S; Draw, A; Mador,
  MJ; El-Solh, A; Grant, BJB. Chest 124(5): 1694-701
  2003 Nov.
Thanks for your attention
                 January 27, 2005

Shared By: