Electronic Supplementary Material to

Incidence and Prognosis of Dysnatremias Present on ICU Admission

Georg-Christian Funk*, Gregor Lindner*, Wilfred Druml,Barbara Metnitz, Christoph Schwarz, Peter Bauer Philipp G.H. Metnitz on behalf of the ASDI Study Group.

Table E1. Unadjusted Associations of Dysnatremia with Patients’co-morbidities and diagnoses upon ICU admission.

Characteristic / Hyponatremia / Normal Sodium / Hypernatremia
Severe / Mild / Borderline / Borderline / Mild / Severe / P
(n = 1,864) / (n = 4,076) / (n = 20,842) / (n = 114,170) / (n = 7,723) / (n = 1,846) / (n = 965)
SAPS II score, median (quartiles) † / 44.0 (33.0-58.0) / 37.0 (27.0-53.0) / 32.0 (22.0-45.0) / 27.0 (19.0-39.0) / 38.0 (26.0-53.0) / 47.0 (34.0-63.0) / 53.0 (41.0-67.0) / <0.0001
SAPS II-predicted mortality, %† / 40.2 / 31.8 / 23.9 / 18.4 / 32.4 / 44.2 / 52.4 / <0.0001
Post-ICU mortality, % / 12.7 / 12.7 / 8.0 / 5.4 / 9.4 / 14.3 / 21.7 / <0.0001
SAPS II-observed-to-expected mortality ratio and 95% confidence interval† / 0.836 (0.792-0.880) / 1.036 (1.001-1.071) / 0.888 (0.869-0.908) / 0.792 (0.782-0.802) / 0.879
(0.854-0.904) / 1.026
(0.986-1.066) / 1.103
(1.054-1.151) / <0.0001
Co-morbidities present upon ICU admission, %
None / 42.7 / 44.0 / 51.7 / 62.5 / 57.6 / 54.1 / 57.4 / <0.0001
Hematologic disease / 1.5 / 1.9 / 1.5 / 1.0 / 1.4 / 2.3 / 1.4 / <0.0001
AIDS / 0.2 / 0.3 / 0.2 / 0.1 / 0.2 / 0.5 / 0.2 / 0.0002
Metastasizing malignancy / 3.3 / 4.7 / 5.5 / 3.9 / 3.1 / 3.2 / 3.5 / <0.0001
Immunosuppression / 3.2 / 3.2 / 2.1 / 1.2 / 1.3 / 2.4 / 2.2 / <0.0001
Chronic renal failure / 10.4 / 11.3 / 8.3 / 6.1 / 7.6 / 8.6 / 8.4 / <0.0001
Chronic respiratory failure / 11.4 / 11.4 / 9.4 / 7.7 / 9.9 / 11.1 / 8.6 / <0.0001
Chronic heart failure / 17.4 / 20.5 / 18.0 / 13.7 / 15.1 / 15.2 / 11.9 / <0.0001
Liver cirrhosis / 12.9 / 7.7 / 4.2 / 2.1 / 3.3 / 4.9 / 4.2 / <0.0001
Alcoholism / 16.0 / 9.7 / 5.8 / 4.2 / 7.6 / 9.6 / 7.4 / <0.0001
Insulin-dependent diabetes / 9.1 / 8.6 / 6.9 / 4.6 / 5.1 / 6.3 / 6.9 / <0.0001
Acute renal failure / 10.8 / 8.7 / 4.4 / 2.0 / 4.4 / 7.0 / 7.2 / <0.0001
Non-metastasizing malignancy / 4.2 / 5.6 / 7.6 / 6.6 / 5.8 / 4.3 / 5.1 / <0.0001
Drug addiction / 1.0 / 1.6 / 1.1 / 1.1 / 1.5 / 1.3 / 1.7 / 0.0008
Main diagnosis upon ICU admission, % / <0.0001
Metabolic disease / 9.6 / 4.5 / 2.7 / 3.9 / 5.7 / 3.5 / 5.9
Respiratory disease / 10.4 / 9.8 / 7.5 / 6.8 / 11.2 / 12.6 / 12.0
Cardiovascular disease / 16.2 / 17.0 / 13.6 / 18.2 / 13.1 / 9.7 / 9.3
Shock / 4.8 / 5.0 / 2.7 / 1.6 / 3.2 / 4.7 / 4.8
Renal disease / 7.3 / 5.3 / 2.6 / 1.1 / 1.6 / 2.0 / 2.3
Neurologic disease / 6.0 / 3.3 / 2.8 / 3.6 / 4.4 / 4.4 / 6.8
Sepsis / 3.4 / 3.5 / 1.9 / 1.0 / 2.4 / 3.4 / 4.6
Trauma (not operated) / 0.9 / 0.9 / 1.0 / 2.2 / 2.9 / 3.0 / 1.7
Gastrointestinal disease / 5.6 / 4.2 / 2.3 / 1.5 / 1.9 / 2.3 / 2.9
Hematologic disease / 0.2 / 0.2 / 0.1 / 0.1 / 0.3 / 0.5 / 0.4
Medical, not otherwise specified / 1.3 / 0.8 / 0.6 / 0.7 / 0.5 / 0.6 / 1.0
Pregnancy / 0.1 / 0.1 / 0.1 / 0.1 / 0.0 / 0.1 / 0.1
Thoracic surgery / 0.7 / 1.0 / 1.5 / 1.0 / 0.5 / 0.4 / 0.3
Cardiovascular surgery / 2.4 / 5.6 / 12.5 / 12.3 / 8.1 / 6.3 / 3.5
Neurosurgery / 2.1 / 2.6 / 3.5 / 3.8 / 4.1 / 3.4 / 5.6
Transplant surgery / 0.5 / 0.7 / 0.8 / 0.6 / 0.7 / 0.5 / 0.2
Trauma surgery / 5.9 / 5.4 / 6.3 / 11.1 / 16.4 / 19.0 / 18.6
Abdominal surgery / 5.5 / 10.2 / 13.4 / 10.5 / 10.7 / 13.1 / 9.7
Surgery, not otherwise specified / 2.6 / 4.1 / 7.9 / 8.9 / 5.4 / 4.4 / 2.9

†SAPS, Simplified Acute Physiology Score. Comparisons between groups were made using ANOVA or the chi-square test.

Table E2. Logistic Regression Model: Influence of Dysnatremia on Hospital Mortality

Odds ratio and 95% Confidence Interval
Hyponatremia
Severe / Na*<125mmol/L / 1.81 (1.56 to 2.10)
Mild / 125 ≤ Na < 130 mmol/L / 1.89 (1.71 to 2.09)
Borderline / 130 ≤ Na < 135 mmol/L / 1.32 (1.25 to 1.39)
Normal sodium (reference group) / 135 ≤ Na ≤ 145 mmol/L
Hypernatremia
Borderline / 145 < Na ≤ 150 mmol/L / 1.48 (1.36 to 1.61)
Mild / 150 < Na ≤ 155 mmol/L / 2.32 (1.98 to 2.73)
Severe / Na >155 mmol/L / 3.64 (2.88 to 4.61)
Sodium-corrected SAPS II score† / 2.15 (2.12 to 2.18)
Interactions of the sodium-corrected SAPS II score with hypo- and hypernatremia
Hyponatremia
Severe / 1.82 (1.69 to 1.96)
Mild / 1.84 (1.75 to 1.95)
Borderline / 1.96 (1.89 to 2.03)
Normal sodium
Hypernatremia
Borderline / 1.95 (1.86 to 2.04)
Mild / 1.91 (1.77 to 2.05)
Severe / 1.84 (1.66 to 2.04)
Type of ICU admission
Unplanned surgery / 0.72 (0.68 to 0.76)
Planned surgery / 0.71 (0.67 to 0.76)
Medical / Reference group
Main diagnosis upon ICU admission
Metabolic disease / 0.27 (0.23 to 0.32)
Respiratory disease / 1.18 (1.11 to 1.26)
Cardiovascular disease / 0.82 (0.78 to 0.87)
Shock / 1.43 (1.30 to 1.58)
Sepsis / 1.29 (1.15 to 1.45)
Gastrointestinal disease / 1.30 (1.16 to 1.46)
Hematologic disease / 1.76 (1.24 to 2.49)
Medical, not otherwise specified / 0.77 (0.62 to 0.96)
Pregnancy / 0.18 (0.03 to 0.96)
Thoracic surgery / 1.41 (1.19 to 1.68)
Neurosurgery / 1.15 (1.05 to 1.26)
Transplant surgery / 0.66 (0.53 to 0.84)
Abdominal surgery / 1.53 (1.44 to 1.63)
Surgery, not otherwise specified / 0.85 (0.78 to 0.93)
Co-morbidities present upon ICU admission
AIDS / 0.49 (0.34 to 0.72)
Chronic heart failure / 1.37 (1.30 to 1.44)
Chronic renal failure / 1.28 (1.21 to 1.35)
Chronic respiratory failure / 1.17 (1.11 to 1.24)
Liver cirrhosis / 2.29 (2.10 to 2.49)
Drug addiction / 0.59 (0.48 to 0.72)
Hematologic disease / 0.85 (0.75 to 0.97)
Immunosuppression / 1.24 (1.11 to 1.40)
Acute renal failure / 1.49 (1.38 to 1.62)
Metastasizing malignancy / 1.16 (1.07 to 1.26)
Non-metastasizing malignancy / 1.14 (1.06 to 1.23)
None / 0.93 (0.88 to 0.98)

*Na, serum sodium;†SAPS, Simplified Acute Physiology Score; area under the receiver operating characteristic curve, 0.868; R2, 0.2505; maximum rescaled R2, 0.4138. The odds ratios of the interaction terms relate to a 10-point increase in the sodium-corrected SAPS II score within each grade of dysnatremia.

Data Quality

To assess the reliability of data collection, we sent an independent observer to each unit to obtain SAPS II data from the clinical charts of a random sample of patients. Variance-component analyses with the random factors “units,” “patients within units,” and “observers within units” were performed (SAS, procedure varcomp) as previously described [1]. To assess completeness of documentation, we also calculated the number of missing variables for the SAPS II score. Additional details have been reported elsewhere [1].

Data quality was satisfactory with respect to both completeness of records and inter-rater variability. The median number of missing variables necessary for the calculation of the SAPS II was 0 (interquartile range, 0 to 2). Inter-rater quality control indicated an excellent grade of agreement: For all tested variables, practically no deviations between the observers were detected, the contribution to the variability being <1%.

Details about the logistic regression model

To exclude case mix factors as possible confounders, the model was adjusted for predefined variables (showing an effect in a univariate unadjusted analysis) affecting hospital mortality: sex, main diagnosis responsible for ICU admission, and co-morbidities on ICU admission, sodium-corrected SAPS II score, and type of ICU admission. Moreover, the code of the ICU where the patient was treated was added to adjust for the effect of different treatment centers. The goodness-of-fit test as described by Hosmer and Lemeshow [2] was used to evaluate the fit of the model.

We included interaction terms between the grade of dysnatremia and the sodium-corrected SAPS II score in the model, to assess the effect of changes in the sodium-corrected SAPS II risk score within the different grades of dysnatremia. The odds ratios of the interaction terms given in Table E2 relate to a 10-point increase in the sodium-corrected SAPS II score within each grade of dysnatremia. The odds ratios of the respective interaction terms (e.g. 0.85 for severe hyponatremia) were multiplied with the odds ratio of the sodium-corrected SAPS II score (2.15) at the mean value of the cohort (32.6). The resulting odds ratio (e.g. 1.82 for severe hyponatremia) displayed in Table E2 can be used to obtain the individual risk of dysnatremic patients with various sodium-corrected SAPS II scores.

Reference List

1. Metnitz PG, Vesely H, Valentin A, Popow C, Hiesmayr M, Lenz K, Krenn CG, Steltzer H (1999) Evaluation of an interdisciplinary data set for national intensive care unit assessment. Crit Care Med 27:1486-1491

2. LEMESHOW STAN, HOSMER DW, Jr. (1982) A REVIEW OF GOODNESS OF FIT STATISTICS FOR USE IN THE DEVELOPMENT OF LOGISTIC REGRESSION MODELS. Am J Epidemiol 115:92-106