Association between hyperkalemia at critical care initiation and mortality.
ABSTRACT To investigate the association between potassium concentration at the initiation of critical care and all-cause mortality.
We performed a retrospective observational study on 39,705 patients, age ≥18 years, who received critical care between 1997 and 2007 in two tertiary care hospitals in Boston, Massachusetts. The exposure of interest was the highest potassium concentration on the day of critical care initiation and categorized a priori as 4.0-4.5, 4.5-5.0, 5.0-5.5, 5.5-6.0, 6.0-6.5, or ≥6.5 mEq/l. Logistic regression examined death by days 30, 90, and 365 post-critical care initiation, and in-hospital mortality. Adjusted odds ratios were estimated by multivariable logistic regression models.
The potassium concentration was a strong predictor of all-cause mortality 30 days following critical care initiation with a significant risk gradient across potassium groups following multivariable adjustment: K = 4.5-5.0 mEq/l OR 1.25 (95 % CI, 1.16-1.35; P < 0.0001); K = 5.0-5.5 mEq/l OR 1.42 (95 % CI, 1.29-1.56; P < 0.0001); K = 5.5-6.0 mEq/l OR 1.67 (95 % CI, 1.47-1.89; P < 0.0001); K = 6.0-6.5 mEq/l OR 1.63 (95 % CI, 1.36-1.95; P < 0.0001); K > 6.5 mEq/l OR 1.72 (95 % CI, 1.49-1.99; P < 0.0001); all relative to patients with K = 4.0-4.5 mEq/l. Similar significant associations post multivariable adjustments are seen with in-hospital mortality and death by days 90 and 365 post-critical care initiation. In patients whose hyperkalemia decreases ≥1 mEq/l in 48 h post-critical care initiation, the association between high potassium levels and mortality is no longer significant.
Our study demonstrates that a patient's potassium level at critical care initiation is robustly associated with the risk of death even at moderate increases above normal.
- SourceAvailable from: Jean-Charles E PreiserEuropean Journal of Intensive Care Medicine 12/2012; · 5.17 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: The study objective was to investigate the association between primary language spoken and all-cause mortality in critically ill patients. We performed a cohort study on 48 581 patients 18 years or older who received critical care between 1997 and 2007 in 2 Boston hospitals. The exposure of interest was primary language spoken determined by the patient or family members who interacted with administrative staff during hospital registration. The primary outcome was 30-day mortality. Associations between language and mortality were estimated by bivariable and multivariable logistic regression models with inclusion of covariate terms thought to plausibly interact with both language and mortality. Adjustment included age, race, sex, Deyo-Charlson index, patient type (medical vs surgical), sepsis, creatinine, hematocrit, white blood count, and number of organs with acute failure. Validation showed that primary language spoken was highly accurate for a statement in the medical record noting the language spoken that matched the assigned language. Patients whose primary language spoken was not English had improved outcomes (odds ratio 30-day mortality, 0.69 [95% confidence interval, 0.60-0.81; P < .001), relative to patients with English as the primary language spoken, fully adjusted. Similar significant associations are seen with death by days 90 and 365 as well as in-hospital mortality. The improved survival in patients with a non-English primary language spoken is not confounded by indicators of severity of disease and is independent of the specific language spoken and neighborhood poverty rate, a proxy for socioeconomic status. There are significant limitations inherent to large database studies that we have acknowledged and addressed with controlling for measured confounding and evaluation of effect modification. In a regional cohort, not speaking English as a primary language is associated with improved outcomes after critical care. Our observations may have clinical relevance and illustrate the intersection of several factors in critical illness outcome including severity of illness, comorbidity, and social and economic factors.Journal of critical care 09/2013; · 2.13 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Red cell distribution width is associated with mortality and bloodstream infection risk in the critically ill. In hospitalized patients with critical illness, it is not known if red cell distribution width can predict subsequent risk of all-cause mortality following hospital discharge. We hypothesized that an increase in red cell distribution width at hospital discharge in patients who survived to discharge following critical care would be associated with increased postdischarge mortality. Two-center observational cohort study SETTING:: All medical and surgical ICUs at the Brigham and Women's Hospital and Massachusetts General Hospital. We studied 43,212 patients, who were 18 years old or older and received critical care between 1997 and 2007 and survived hospitalization. None. The exposure of interest was red cell distribution width within 24 hours of hospital discharge and categorized a priori in quintiles as less than or equal to 13.3%, 13.3-14.0%, 14.0-14.7%, 14.7-15.8%, and more than 15.8%. The primary outcome was all-cause mortality in the 30 days following hospital discharge. Secondary outcomes included 90-day and 365-day mortality following hospital discharge. Mortality was determined using the U.S. Social Security Administration Death Master File, and 365-day follow-up was present in all cohort patients. Adjusted odds ratios were estimated by multivariable logistic regression models with inclusion of covariate terms thought to plausibly interact with both red cell distribution width and mortality. Adjustment included age, race, gender, Deyo-Charlson Index, patient type (medical vs surgical), sepsis, and number of organs with acute failure. In patients who received critical care and survived hospitalization, the discharge red cell distribution width was a robust predictor of all-cause mortality and remained so following multivariable adjustment. Patients with a discharge red cell distribution width of 14.0-14.7%, 14.7-15.8%, and more than 15.8% have an odds ratio for mortality in the 30 days following hospital discharge of 2.86 (95% CI, 2.25-3.62), 4.57 (95% CI, 3.66-5.72), and 8.80 (95% CI, 7.15-10.83), respectively, all relative to patients with a discharge red cell distribution width less than or equal to 13.3%. Following multivariable adjustment, patients with a discharge red cell distribution width of 14.0-14.7%, 14.7-15.8%, and more than 15.8% have an odds ratio for mortality in the 30 days following hospital discharge of 1.63 (95% CI, 1.27-2.07), 2.36 (95% CI, 1.87-2.97), and 4.18 (95% CI, 3.36-5.20), respectively, all relative to patients with a discharge red cell distribution width less than or equal to 13.3%. Similar significant robust associations post multivariable adjustments are seen with death by days 90 and 365 postdischarge. Estimating the receiver-operating characteristic area under the curve shows that discharge red cell distribution width has moderate discriminative power for mortality 30 days following hospital discharge (area under the curve = 0.70; SE 0.006; 95% CI, 0.69-0.71; p < 0.0001). In patients treated with critical care who survive hospitalization, an elevated red cell distribution width at the time of discharge is a robust predictor of subsequent all-cause patient mortality. Increased discharge red cell distribution width likely reflects the presence of proinflammatory state, oxidative stress, arterial underfilling, or a combination, thereof which may explain the observed impact on patient survival following discharge. Elevated red cell distribution width at hospital discharge may identify ICU survivors who are at risk for adverse outcomes following hospital discharge.Critical care medicine 01/2014; · 6.37 Impact Factor