Efigure 1. Consort Type Diagram Showing Enrollment and Follow-Up of Patients

Decreased Respiratory Rate Variability During Mechanical Ventilation is Associated with Increased Mortality

Gutierrez et al

On-Line Only Supplement

eFigure 1. Consort type diagram showing enrollment and follow-up of patients.

eTable 1 - MedicalDiagnoses and Surgical Procedures at Enrollment

eTable 2 - Ethnicity and Risk Factors

eTable 3 - Laboratory Values at the Time of Enrollment

eTable 4 - Time Averaged Hemodynamic Variables and Blood Gases Data

eTable 5 - Primary and Secondary Outcomes

eTable6 - Logistic Regression Analysis

eTable7 - Intravenous Medications

Sensitivity Analysis for H1/DC Boundary

eFigure1. Consort type diagram showing enrollment and follow-up of patients according to study group.

eTable 1
Medical Diagnoses and Surgical Procedures at Enrollment
RRVHigh (n = 47) / RRVLow (n = 131)
Medical Diagnoses (%) ¶† / (n = 36) / (n=94) / p value
Pneumonia / 19 / 12 / 0.25
Acute CVA / 19 / 14 / 0.43
Sepsis / 17 / 16 / 0.92
AMI and Cardiogenic Shock / 14 / 21 / 0.34
GI Bleeding / 11 / 9 / 0.65
ARDS / 3 / 6 / 0.42
Seizures / 3 / 3 / 0.90
Surgical Procedures (%) ¶ / (n = 15) / (n=40)
Abdominal / 13 / 40 / 0.06
Cardiothoracic ** / 33 / 5 / 0.005
Neurosurgery / 13 / 8 / 0.50
Trauma / 27 / 35 / 0.56
Vascular / 7 / 3 / 0.46

RRVLow = Patients with low respiratory rate variability.

RRVHigh = Patients with high respiratory rate variability.

¶ Percent of patients in the group.

† The sum of all patients with medical and surgical diagnoses exceeds the total number of patients. Some patients had both medical and surgical diagnoses as a reason for ICU admission, i.e a postoperative patient with sepsis.

CVA = Cerebro-vascular accident; AMI = Acute myocardial infarction; GI = Gastrointestinal; ARDS = Acute respiratory distress syndrome.

* p<0.05; ** p < 0.01. Unless stated, figures shown as median [interquartile range].

eTable2
Ethnicity and Risk Factors
RRVHigh(n = 47) / RRVLow(n=131) / p value
Ethnicity (%)¶
African American / 51.1 / 58.0 / 0.41
Caucasian / 38.3 / 34.3 / 0.63
Hispanic / 6.4 / 2.3 / 0.18
Asian / 2.1 / 3.1 / 0.74
Multiracial / 2.1 / 2.3 / 0.95
Risk Factors (%)¶
Smoker / 19 / 21 / 0.83
ETOH / 6 / 11 / 0.39
HIV / 9 / 2 / 0.06
IVDA / 0 / 3 / 0.23
Metastatic cancer / 2 / 4 / 0.58
Hematologic malignancy / 0 / 1 / 0.55

RRVLow = Patients with low breathing variability.

RRVHigh = Patients with high breathing variability.

¶ % of patients in each group

ETOH = alcohol abuse

IVDA = intravenous drug abuse

HIV = Human immunodeficiency virus positive.

eTable3
Laboratory Valuesat the Time of Enrollment†
RRVHigh / RRVLow
n / n / p value
pHa / 47 / 7.39 [7.34 - 7.44] / 131 / 7.37 [7.32 - 7.43] / 0.36
PaCO2 (mmHg) / 47 / 37 [33 – 46] / 131 / 38 [32 – 44] / 0.99
PaO2 (mmHg) / 47 / 159 [112 – 197] / 131 / 135 [90 – 198] / 0.95
WBC x 10-3 / 46 / 11.1 [7.8 – 13.1] / 131 / 10.4 [7.4 – 14.7] / 0.44
Hemoglobin (g.dL-1) / 46 / 10 [9 – 11] / 131 / 11 [9 – 12] / 0.39
Platelets x 10-3 / 45 / 192 [133 – 282] / 131 / 204 [135 – 275] / 0.84
Sodium (mmol.L-1) / 47 / 139 [137 – 142] / 131 / 139 [136 – 142] / 0.89
Potassium(mmol.L-1) / 47 / 3.9 [3.5 - 4.4] / 131 / 4.1 [3.7 - 4.5] / 0.20
Chloride(mmol.L-1) / 47 / 107 [102 – 109] / 131 / 106 [102 – 110] / 0.92
CO2 (mmol.L-1) / 47 / 24 [21 – 27] / 131 / 23 [20 – 27] / 0.75
BUN (mg.dL-1) / 47 / 19 [11 – 27] / 131 / 16 [12 – 26] / 0.85
Creatinine (mg.dL-1) / 47 / 0.9 [0.8 - 1.4] / 131 / 1.0 [0.7 - 1.4] / 0.68
Magnesium (mEq.L-1) / 39 / 1.8 [1.6 – 2.1] / 117 / 1.7 [1.5 - 2.0] / 0.59
Calcium (mg.dL-1) / 42 / 7.9 [7.4 - 8.7] / 124 / 7.8 [7.2 - 8.4] / 0.19
Phosphorus (mg.dL-1) / 29 / 3.5 [2.7 – 4.7] / 76 / 3.6 [3.0 - 4.5] / 0.45
INR / 28 / 1.2 [1.1 - 1.4] / 88 / 1.3 [1.1 - 1.5] / 0.26
PTT (s) / 26 / 35 [30 – 45] / 84 / 35 [29 – 44] / 0.48
Glucose (mg.dL-1) / 42 / 123 [104 – 161] / 130 / 130 [107 – 159] / 0.97
Lactate (mmol.dL-1) / 39 / 1.7 [1.3 - 3.4] / 114 / 2.0 [1.3 - 3.2] / 0.57

RRVLow = Patients with low breathing variability.

RRVHigh = Patients with high breathing variability.

† Data includes only those patients who had laboratory tests performed at the time of ICU admission.

n = Number of samples; pHa = Arterial pH; PaCO2 = Partial pressure of arterial CO2; PaO2 =Partial pressure of arterial O2; WBC= white blood cells; BUN= Blood urea nitrogen; INR= International normalized ratio; PTT= Partial thromboplastin time.

Figures shown as median [interquartile range].

eTable4
Time Averaged Hemodynamic Variables and Blood Gases Data
RRVHigh (n = 47) / RRVLow (n=131) / p value
Hemodynamic Variables¶
Heart Rate (bpm) / 87.5 [78.8 – 95.8] / 86.1 [75.7 – 96.1] / 0.25
Mean Arterial Pressure (mmHg) / 84.4 [79.4 – 93.6] / 82.9 [77.2 – 91.7] / 0.44
Body Temperature (OC) / 37.1 [36.8 - 37.4] / 37.1 [36.6 - 37.4] / 0.81
Cardiac Output (L.min-1)† / 6.5 [4.9 – 7.9] / 6.6 [5.3 – 7.7] / 0.83
RRVHigh (n = 47) / RRVLow (n=131) / p value
Blood Gases Data¶ / (n = 38) / (n=119)
pHa / 7.41 [7.39 - 7.44] / 7.41 [7.36 - 7.43] / 0.82
PaCO2 (mmHg) / 37 [34 - 41] / 38 [34 - 42] / 0.42
PaO2 (mmHg) / 139 [123 - 162] / 137 [106 - 164] / 0.98
SaO2 (%) / 98.5 [97.1 - 99.4] / 98.3 [96.4 - 99.3] / 0.52
PaO2/FIO2 (mmHg) / 360 [275 – 390] / 307 [229 – 397] / 0.56

RRVLow = Low respiratory rate variability. RRVHigh = High respiratory rate variability.

¶ Mean values computed during the time patients were monitored.

† Cardiac output recorded with an arterial pulse-wave contour device (n = 97; RRVHigh n=26;RRVLow n=71 ).

pHa = Arterial pH

PaCO2 = Partial pressure of arterial CO2

PaO2 =Partial pressure of arterial O2;

SaO2 = Arterial O2 saturation.

* p<0.05; ** p < 0.01. Unless stated, figures shown as median [interquartile range].

eTable5
Primary and Secondary Outcomes
All Cause Mortality Rates† / RRVHigh
(n = 47) / RRVLow
(n = 129) ¶ / p value
At 28 days * / 10.6% / 28.7% / 0.013
(n = 47) / (n = 125) §
At 60 days * / 19.1% / 32.8% / 0.046
At 90 days / 23.4% / 33.6% / 0.13
At 180 days / 29.8% / 37.6% / 0.15
Other Outcomes / (n = 47) / (n=131)
Days on Mechanical Ventilation / 7.0 [3.0 – 12.5] / 5.0 [3.0 – 10.0] / 0.27
Ventilator Free Days / 21.0 [15.5 – 25.0] / 23.0 [19.0 – 25.0] / 0.27
ICU LOS (Days) / 11.0 [5.5 – 19.5] / 10.0 [5.0 - 19.0] / 0.66
Hospital LOS (Days) / 23.0 [12.0 – 32.0] / 18.0 [10.0 – 32.0] / 0.29
Barotrauma / 0.0% / 3.8% / 0.17
VAP / 21.3% / 19.8% / 0.83
Sepsis / 2.1% / 6.1% / 0.29
Tracheostomy / 14.9% / 7.6% / 0.15
Failure to Wean / 19.1% / 13.7% / 0.38
Self Extubation / 2.1% / 6.1% / 0.29
Transfer to a VLTCF / 8.5% / 3.1% / 0.12

RRVLow = Low respiratory rate variability. RRVHigh = High respiratory rate variability.

† Computed from the day of enrollment.

¶ Two patients in this group were lost to follow-up prior to 28 days and were excluded when calculating mortality rates.

§ Six patients in this group were lost to follow-up prior to 60 days and were excluded when calculating mortality rates.

LOS = Length of stay; VAP= Ventilator-associated pneumonia (pneumonia arising more than 48-72 hours after endotracheal intubation); Failure to wean = Reintubation within 48 hours of weaning; Barotrauma = Newly developed pneumothorax, pneumomediastinum or subcutaneous emphysema; VLTCF = Ventilator Long Term Care Facility.

* p<0.05. Unless stated, figures shown as median [interquartile range].

eTable6

Logistic Regression Analysis

Significant Variables / O.R. / 95% Confidence Interval / p value
RRVLow(H1/DC ≥ 40%) ** / 6.08 / 1.67 / 22.20 / 0.006
i.v. vasopressin ** / 4.59 / 1.73 / 12.15 / 0.002
SAPS II ** / 1.08 / 1.03 / 1.13 / 0.001
Not Significant Variables
Age / 1.01 / 0.98 / 1.04 / 0.36
Average Mean Pressure / 1.04 / 0.97 / 1.11 / 0.29
Average Peak Pressure / 0.94 / 0.86 / 1.02 / 0.29
Body Mass Index / 1.01 / 0.95 / 1.08 / 0.75
i.v. cisatracurium / 0.998 / 0.27 / 3.72 / 0.99
Mean Respiratory Rate / 1.04 / 0.96 / 1.14 / 0.35
Positive End Expiratory Pressure / 0.86 / 0.68 / 1.09 / 0.21

The covariates H1/DC, SAPS II and age were chosen a priori, since they were expected to influence mortality. The other variables were chosen a posteriori, since they were significantly different or, in the case of vasopressin, nearly statistically significantly when comparing low and high RRV groups.

RRVLow = Low breathing variability.

SAPS II = Simplified Acute Physiology Score

O.R = Odds ratio.

* p 0.05 ** p 0.01.

eTable7
Intravenous Medications
RRVHigh (n = 47) / RRVLow (n = 131)
Patients Used (%) / Time Used (%) / Mean
Dose / Patients Used (%) / Time Used (%) / Mean
Dose / p value
Neuromuscular Blockers
Cisatracurium (mcg/kg/min) * / 2.1 / 1.3 / 2.3 / 16.8 / 12.7 / 2.1 / 0.011
Sedatives
Propofol (mcg/kg/min) / 72.3 / 30.2 / 30.5 / 74.8 / 27.9 / 29.5 / 0.74
Fentanyl (mcg/kg/h) / 57.4 / 25.3 / 1.4 / 66.4 / 31.3 / 1.9 / 0.28
Dexmedetomidine (mcg/kg/h) / 19.1 / 8.5 / 1.5 / 14.5 / 14.3 / 1.6 / 0.46
Lorazepam (mg/h) / 0.0 / 6.1 / 12.4 / 3.1 / 0.09
Midazolam (mg/h) / 0.0 / 3.1 / 4.9 / 1.3 / 0.23
Vasoactive Agents
Norepinephrine (mcg/kg/min) / 36.2 / 12.9 / 0.1 / 41.2 / 15.2 / 0.1 / 0.55
Phenylephrine (mcg/kg/min) / 23.4 / 11.9 / 0.6 / 14.5 / 9.6 / 0.6 / 0.17
Vasopressin (units/min) / 12.8 / 19.7 / 0.0 / 26.0 / 19.2 / 0.0 / 0.07
Dopamine (mcg/kg/min) / 4.3 / 26.3 / 4.0 / 9.9 / 14.1 / 5.8 / 0.24
Dobutamine (mcg/kg/min) / 4.3 / 7.6 / 15.9 / 9.2 / 11.5 / 4.2 / 0.29
Epinephrine (mcg/kg/min) / 2.1 / 21.0 / 0.1 / 8.4 / 22.3 / 0.1 / 0.15
Cardiac Medications
Amiodarone (mg/min) / 12.8 / 9.6 / 0.6 / 13.7 / 15.3 / 0.6 / 0.87
Nicardapine (mg/h) / 12.8 / 12.6 / 8.7 / 13.7 / 10.9 / 8.4 / 0.87
Aminocaproic acid (g/h) / 2.1 / 3.2 / 3.7 / 0.8 / 7.3 / 2.0 / 0.45
Milrinone (mcg/kg/min) / 2.1 / 18.6 / 0.8 / 3.8 / 9.2 / 0.4 / 0.59
Other
Furosemide (mg/h) / 23.4 / 11.3 / 4.2 / 13.0 / 9.6 / 6.4 / 0.10
Insulin (units/h) / 19.1 / 7.9 / 6.5 / 17.6 / 13.4 / 6.4 / 0.81
Heparin (units/h) / 8.5 / 23.7 / 1015 / 10.7 / 18.7 / 1068 / 0.68
Octreotide (mcg/h) / 4.3 / 34.3 / 42.6 / 1.5 / 22.3 / 50.0 / 0.29
Pantoprazole (mg/h) / 2.1 / 2.9 / 8.0 / 5.3 / 32.9 / 8.0 / 0.37
Drotrecogin alfa (mcg/kg/h) / 0.0 / 4.6 / 17.1 / 24.0 / 0.14

RRVLow = Low respiratory rate variability. RRVHigh = High respiratory rate variability.

% Time Used = Percentage of monitored time during which the i.v. medication was administered.

p value = Pearson's chi-square comparing percentage of patients receiving the i.v. medication.

* p < 0.05

Sensitivity Analysis for H1/DC Boundary

We performed a sensitivity analysis based on the cohort's 28-day mortality rate, with the aim of defining the optimal H1/DC boundary between regular (RRVLow) and irregular (RRVHigh) breathing patterns. We compared increasing values of H1/DC ranging from 25% to 50% at intervals of 0.01% (H1/DCboundary ) to each patient's time-averaged H1/DC computed for the duration of ventilatory support (H1/DCmean). A patient was labeled RRVHigh if H1/DCmean < H1/DCboundary andRRVLow if H1/DCmean ≥ H1/DCboundary. This categorization resulted in X patients classified as RRVHigh and Y patients as RRVLow for each chosen H1/DCboundary. We next determined the 28-day mortality rate for the X and Y patients (XDead and YDead) and developed the following Latin Square:

RRVHigh / XDead / XAlive / XTotal
RRVLow / YDead / YAlive / YTotal

Applying Pearson's chi-square test resulted in a p value for each H1/DCboundary. The distribution of p values according to H1/DCboundary is shown below, where the ordinate scale is probability and dashed lines are drawn at p values of 0.05 and 0.01:

According to this graph, any value for H1/DCboundary in the range of 34.9% to 42.3% would have yielded a statistically significant difference between the RRVLow and RRVHigh groups at p < 0.05. The p value nadir of 0.003 corresponds to H1/DCboundary = 41.6%.

1