This retrospective study was based on a historical cohort of patients older than 18 years who required cf-LVAD therapy (HVAD, HeartWare Inc) between 2009 and 2015 in our institution. The study information was extracted from a local database and was supplemented by medical record review. The research was conducted in accordance with the principles of the Declaration of Helsinki and the the study was approved by the ethics committee of the center.

All patients with concomitant cardiac surgery were excluded from the study. The NRI was determined using the modified formula as NRI = (1.519 × serum albumin [g/L]) + 41.7 × (current body weight [kg]/IBW [kg]). The IBW was estimated using the Lorentz formulae for men, namely, IBW = height (cm) – 100 – (height [cm] – 150)/4, and for women, namely, IBW = height (cm) – 100 – (height [cm] – 150)/2.5.

As described in previous studies,13,15 a value of 1 was assigned to the term (current body weight [kg]/IBW [kg]) when the result was ≥ 1. Serum albumin values and body weight were evaluated at the closest time point before cf-LVAD implantation.

Right-sided HF was defined as proposed by the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS): elevated central venous pressure (>16mmHg) with depressed cardiac index (<2 L/min/m2) in the absence of elevated pulmonary wedge pressure > 18mmHg requiring implantation of a right ventricular assist device or prolonged (> 1 week) administration of nitric oxide or inotropic therapy. Respiratory failure was defined as pulmonary insufficiency requiring intubation and ventilation for 96hours or more at any time during the postoperative stay due to a blood oxygen saturation < 96% despite respiratory assistance with a fraction of inspired oxygen ≥ 0.50. Major bleeding was defined as an episode of internal or external bleeding causing 1 or more of the following: death, reoperation, hospitalization, or red blood cell transfusion (the first 7 days after implantation: ≥ 4 packed red blood cells in 24hours; ≥ 7 days after implantation: any transfusion of packed red blood cells). Acute renal failure was defined by a new need for dialysis or an increase in serum creatinine to 3 times higher than its baseline value or to > 5mg/dL. Postoperative infection was defined as any infection proven by microbiological isolation requiring treatment with intravenous antibiotics during the postoperative stay. Vital status at 1 year of the cf-LVAD was determined in all patients. Death from any cause during this period was the main study end point.

In this work, qualitative variables are reported as frequency and percentage and continuous variables as mean ± standard deviation when normally distributed or as median [interquartile range] when not. For comparison of the initial characteristics between the different NRI categories, the Pearson chi-square test was used for qualitative variables. The ANOVA test with a first-order polynomial contrast was applied for quantitative variables.

A Cox proportional hazards model was applied to explore the relationship between the variables and the end point after cf-LVAD implantation. The variables were consecutively tested in a backward stepwise Cox multiple regression model to determine the independent predictors of death. The variables used to calculate the NRI were not included in the multivariable models. The NRI was analyzed as a continuous variable. Statistical significance was set at a P value of < .05. Only variables with P < .05 in the univariable model were included in the multivariable analysis. Finally, the Kaplan-Meier method was used to generate survival curves during the first year after cf-LVAD implantation according to nutritional risk. The survival curves were assessed with the log-rank test. The level of statistical significance was set at P < .05 for all comparisons. Statistical analysis was performed with SPSS 20.0.

From January 2009 to December 2015, 279 patients older than 18 years underwent cf-LVAD therapy (HVAD, HeartWare Inc). Before implantation, 15 patients (5.4%) had severe nutritional risk (NRI < 83.5), 60 (21.5%) had moderate nutritional risk (83.5 ≤ NRI < 97.5), and 26 (9.3%) had mild nutritional risk (97.5 ≤ NRI < 100); 178 (63.8%) showed no evidence of malnutrition (NRI ≥ 100). The average preoperative NRI was 105.3 ± 13.6.

The initial characteristics of the subgroups according to NRI quartile are summarized in Table 1. Patients with a lower preoperative NRI had a lower body mass index (P < .001) and body surface area (P < .001), lower serum albumin (P < .001), and smaller ventricles (P = .007). Although the differences were not statistically significant, there were higher percentages of patients with INTERMACS 1 in the group with mild nutritional risk and of patients with INTERMACS 2 in the group with severe nutritional risk. The group with high nutritional risk also contained fewer patients with a history of cardiac surgery (P = .511). A direct proportional relationship was also observed between the number of platelets and the NRI.

The adverse clinical events occurring during the postoperative period after the cf-LVAD therapy are summarized in Table 2. Patients were divided into 4 subgroups according to their preoperative NRI. The incidences of right ventricular failure (P = .041), respiratory failure (P = .045), and infections (P = .047) were inversely correlated with the respective preoperative NRI values. Although the difference was not significant, patients at risk of malnutrition had a shorter stay in the intensive care unit (P = .093) than those without risk. No differences were observed in the incidences of acute renal failure, stroke, pump thrombosis, or postoperative bleeding.

Multivariable logistic regression analysis identified preoperative NRI as a significant independent predictor of right ventricular failure after cf-LVAD implantation (adjusted odds ratio [aOR] per 1 unit = 0.963, 95% confidence interval [95%CI], 0.934-0.992, P = .014), respiratory failure (aOR per 1 unit = 0.961, 95%CI, 0.936-0.987, P = .004), and postoperative infection (aOR per 1 unit = 0.968, 95%CI, 0.946-0.991, P = .007) (Table 3).

During the first year of follow-up after cf-LVAD therapy, 65 patients died (23.3%). Considering the preoperative nutritional risk (severe, moderate, mild, or absent), the mortality rates of the first year were 53.3%, 31.7%, 23.1%, and 18.0% (P = .005), respectively. The main causes of death during the first year were infection (30.8%), multiorgan failure (26.2%), right ventricular failure (13.2%), and bleeding (12.3%). The distribution of the cause of death according to NRI is shown in Figure 1.

Figure 1.

Causes of death during the first postimplantation year according to NRI (percentage of each cause of death within each NRI group). CHF, chronic heart failure; MOF, multiorgan failure; NRI, nutritional risk index; RVF, right ventricular failure.

(0.26MB).

Preoperative NRI was identified in Cox multivariate regression analysis as an independent predictor of lower risk of death from any cause during the first year after cf-LVAD therapy (adjusted hazard ratio per 1 unit = 0.961, 95%CI, 0.941-0.981, P < .001). INTERMACS classification, a history of cardiac surgery, age, and baseline creatinine were also independent predictors of mortality during the first year after implantation (Table 3). Kaplan-Meier survival curves according to NRI subgroup based on malnutrition risk are shown in Figure 2 (log-rank test, P < .001). The ROC curves of the different factors associated with nutrition–NRI, albumin, percentage of ideal weight, and body mass index–are compared in Figure 3; as evident in the figure, the NRI occupies a larger area on the curve.

Figure 2.

Survival curves the first year after implantation of a continuous-flow left ventricular assist device. NRI, nutritional risk index.

(0.15MB).

Figure 3.

Receiver operating characteristic curves for NRI, serum albumin, IBW percentage, and BMI variables. BMI, body mass index; IBW, ideal body weight; NRI, nutritional risk index; ROC, receiver operating characteristic.

(0.1MB).

This study was conducted in a single large tertiary referral center; therefore, the results may not be generalizable. Although retrospective in nature, the study was performed in prospectively compiled data, which involves inherent limitations, such as an inability to adjust for unknown confounding factors and a referral bias. Because the single-center setting of the study weakens the external validity of its results, the conclusions cannot be directly extrapolated to other populations. Although data were collected using the last available blood test before cf-LVAD implantation, the retrospective use of these data cannot exclude the possibility that these values could have varied in the time from the preoperative determination of patients’ albumin and body weight to the implantation.