Abstract
Objectives
Preoperative hypoalbuminaemia is associated with adverse outcome, including increased postoperative mortality in cardiovascular surgery, neurosurgery, trauma and orthopaedic surgery. However, much less is known about the association between preoperative serum albumin and clinical outcomes after liver surgery. In this study, we sought to determine whether hypoalbuminaemia before partial hepatectomy is associated with a worse postoperative outcome.
Design
Observational study.
Setting
University Medical Centre in Germany.
Participants
We analysed 154 patients enrolled in the perioperative PHYsostigmine prophylaxis for liver resection patients at risk for DELIrium and postOperative cognitive dysfunction (PHYDELIO) trial with a preoperative serum albumin assessment. Hypoalbuminaemia was defined as serum albumin <35 g/L. Subgroups classified as hypoalbuminaemia and non-hypoalbuminaemia consisted of 32 (20.8%) and 122 (79.2%) patients, respectively.
Outcome measures
The outcome parameters of interest were postoperative complications according to Clavien (moderate: I, II; major: ≥III), length of intensive care unit (ICU) stay, length of hospital stay and survival rates 1 year after surgery.
Results
Preoperative hypoalbuminaemia was associated with the occurrence of major postoperative complications (OR 3.051 (95% CI 1.197 to 7.775); p=0.019) after adjusting for age, sex, randomisation, American Society of Anesthesiologists physical status, preoperative diagnosis and Child-Pugh class. Both ICU and hospital lengths of stay were significantly prolonged in patients with preoperative hypoalbuminaemia (OR 2.573 (95% CI 1.015 to 6.524); p=0.047 and OR 1.296 (95% CI 0.254 to 3.009); p=0.012, respectively). One-year survival was comparable between patients with and without hypoalbuminaemia.
Conclusions
We found that low serum albumin before surgery was associated with a worse short-term outcome after partial hepatectomy, which strengthens the prognostic value of serum albumin in the setting of liver surgery.
Trial registration numbers
ISRCTN18978802 and EudraCT 2008-007237-47.
Keywords: Hepatobiliary disease, Hepatobiliary surgery, Adult intensive & critical care, NUTRITION & DIETETICS, Anaesthesia in oncology, Clinical Trial
STRENGTHS AND LIMITATIONS OF THIS STUDY.
First study investigating the association of preoperative hypoalbuminaemia and postoperative outcome in a mixed liver surgical cohort independent of the pre-existing condition.
High data quality due to prospectively assessed data.
Postoperative complications were systematically classified according to Clavien.
No adjustment for liver cirrhosis was made.
Introduction
The hepatic protein albumin is acknowledged as a marker of nutritional state, inflammation, hepatic function and body water balance.1–4 Low serum albumin concentrations occur in states of liver and renal dysfunction, sepsis and heart failure, resulting in severe ascites and coagulopathy.1 2 5–8 Experimental studies have demonstrated that serum albumin exerts anti-inflammatory effects, while hypoproteinaemia has been related to impaired activation and increased apoptosis of macrophages.7 9 10 Impaired immune function is associated with adverse outcomes, particularly in surgical patients.11 Preoperative hypoalbuminaemia is related to increased postoperative morbidity and mortality after cardiovascular surgery,12–17 neurosurgery18 19 and orthopaedic surgery.12–21 However, little is known about the association between hypoalbuminaemia and the postoperative outcome in liver surgery. Previous studies have used serum albumin to calculate various scores rather than using it as an independent outcome marker.22–24 As previous research has primarily been restricted to patients with hepatocellular carcinoma (HCC), the aim of this subgroup investigation was to evaluate whether preoperative serum albumin as a single parameter is related to the postoperative outcome independent of the preoperative diagnosis in patients undergoing partial hepatectomy.
Material and methods
This research is being reported in line with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for reporting observational studies.25
Patients
This is a subgroup investigation of the randomised placebo-controlled Perioperative PHYsostigmine prophylaxis for liver resection patients at risk for DELIrium and postOperative cognitive dysfunction (PHYDELIO) trial.26 The initial PHYDELIO trial recruited 261 patients undergoing open or laparoscopic partial hepatectomy. Patients were eligible if they were at least 18 years of age and not coenrolled in another clinical trial. Written informed consent was obtained from all participants. Between August 2009 and March 2016, patients were randomly assigned to receive either physostigmine (Anticholium, Köhler, Germany) or placebo for 24 hours. The primary outcome measures of the PHYDELIO trial were postoperative delirium and postoperative cognitive dysfunction.
All patients received guideline-based anaesthesiological and surgical treatment according to our standard operating procedures.27
For the purpose of this subgroup investigation, patients with no assessment of preoperative serum albumin were excluded. Furthermore, patients with suspected pseudohypoalbuminaemia were excluded from the analysis. Pseudohypoalbuminaemia was classified as hypoalbuminaemia in the presence of haematocrit (Hct) greater than 10% below the lower limit reference value, reflecting values below the 5th percentile. The Hct lower limit reference values were set according to the local reference laboratory Labor Berlin - Charité Vivantes GmbH and were as follows: men ≤65 years: 40%; women ≤65 years: 36%; men >65 years: 37%; women >65 years: 35%.28
Serum albumin concentrations and definition of hypoalbuminaemia
Preoperative serum albumin was extracted from the patients’ electronic laboratory records. If multiple assessments were available, then the measurement at the closest date to surgery was considered. Hypoalbuminaemia was defined as a serum albumin concentration of <35 g/L according to our reference laboratory.28
Study endpoints and data acquisition
The outcome measures of interest in this analysis were postoperative complications according to the classification by Clavien-Dindo,29 lengths of intensive care unit (ICU) and hospital stay and 1-year survival after surgery. Research physicians and trained research team members assessed all the outcome measures and baseline characteristics. All the variables were prospectively documented in paper-based case report forms and an electronic study database and were reviewed by an internal research associate in accordance with the Good Clinical Practice guideline. Preoperatively acquired data included patients’ demographic characteristics, surgical diagnosis (ie, malignant, metastases, benign, miscellaneous), medical comorbidities, physical status according to the American Society of Anesthesiologists (ASA), presence of ascites or alcohol use disorders using the Alcohol Use Disorders Identification Test questionnaire. The type and duration of surgery were also recorded as was evidence of ascites and serum albumin substitution in the postoperative period.
Statistical analysis
The results are expressed according to their scaling as medians (25%, 75% quartiles) or frequencies (%). Differences between groups regarding continuous parameters were tested using non-parametric exact Mann-Whitney U tests for independent groups. For frequencies, the χ2 test was applied. For parameters with ordered categories, the linear-by-linear association test was used. The number of missing data are given below each variable.
Associations between the outcome parameters and relevant clinical parameters were tested using binary logistic regression models with adjustment for possible confounders. For this purpose, length of ICU stay and length of stay (LOS) as dependent variables were dichotomised using their respective medians as cutoffs. Postoperative morbidity was subsequently classified as moderate (Clavien grades I–II) and severe (Clavien grade ≥III), with the latter being the dependent variable. Odds ratios (OR) with 95% confidence intervals (CI) and the corresponding p values were calculated. For regression models with hospital LOS as dependent variable, ORs and 95% CIs were calculated based on 1000 bootstrap samples for each covariate to gain more robustness in the estimates. Mortality rates in both groups were estimated according to Kaplan-Meier analysis and are reported with the 95% CI. Survival until 12 months was compared between treatment groups using the log-rank test. Cox regression was applied to test for differences in survival with adjustment for confounding factors. Hazard ratios (HR) with 95% CIs and the corresponding p values were calculated. A two-tailed p<0.05 was considered statistically significant. All p values were from exploratory data analysis without adjustment for multiple comparisons. No sample size calculation was performed a priori for this subgroup investigation. All the calculations were performed using IBM SPSS Statistics, V.23.
Patient and public involvement statement
None.
Results
Out of 261 patients enrolled in the PHYDELIO trial, 11 patients had to be excluded because no liver resection was performed, 4 patients had to be excluded because they had pseudohypoalbuminaemia, 91 patients had to be excluded because no preoperative serum albumin assessment was performed and 1 patient had to be excluded because no haematocrit measurement was performed. In all patients, preoperative serum albumin was assessed either 1 day before surgery or at the day of surgery. A total of 154 patients constituted the study population of this investigation. In total, seven patients underwent reresection of the liver, of whom five had new colorectal liver metastases and two suffered from tumour recurrence of their HCC. Figure 1 shows a flow diagram for the inclusion of patients. Hypoalbuminaemia (<35 g/L) was detected in 32 patients (20.8%). Patients with hypoalbuminaemia had a lower haematocrit level, higher levels of transaminases, alkaline phosphatase (AP) and gamma-glutamyl transferase (gGT) preoperatively. Their preoperative liver function was more likely to be classified as Child B and they underwent longer surgery. Other demographic and medical baseline characteristics were comparable between groups (table 1).
Figure 1.
Consolidated Standard of Reporting Trials (CONSORT) flow chart.
Table 1.
Baseline characteristics
| Full sample N=154 | No hypoalbuminaemia n=122 | Hypoalbuminaemia n=32 | P value | |
| Randomisation | 0.694† | |||
| Placebo n (%) | 81 (52.6) | 63 (51.6) | 18 (56.3) | |
| Physostigmine n (%) | 73 (47.4) | 59 (48.4) | 14 (43.8) | |
| Age, years | 60 (52, 67) | 60 (52, 67) | 62 (53, 69) | 0.462* |
| Sex | 0.423† | |||
| Female n (%) | 62 (40.3) | 47 (38.5) | 15 (46.9) | |
| Male n (%) | 92 (59.7) | 75 (61.5) | 17 (53.1) | |
| BMI, kg/m2 | 25.1 (22.8, 29.0) | 25.3 (23.1, 28.8) | 24.6 (22.0, 29.5) | 0.362* |
| ASA | 0.531‡ | |||
| I, n (%) | 11 (7.1) | 8 (6.6) | 3 (9.4) | |
| II, n (%) | 118 (76.6) | 93 (76.2) | 25 (78.1) | |
| III, n (%) | 25 (16.2) | 21 (17.2) | 4 (12.5) | |
| Charlson Comorbidity Index | 5.5 (2.0, 6.0) | 5.5 (2.0, 6.0) | 5.5 (2.0, 6.0) | 0.939* |
| NYHA | 36 (23.4) | 29 (23.8) | 7 (21.9) | >0.999† |
| Arterial hypertension n (%) | 65 (42.2) | 50 (41.0) | 15 (46.9) | 0.688† |
| Diabetes n (%) | 18 (11.7) | 15 (12.3) | 3 (9.4) | 0.767† |
| AUDIT n (%) | 13 (8.4) | 12 (9.8) | 1 (3.1) | 0.304† |
| Chemotherapy presurgery n (%) | 36 (23.4) | 29 (23.8) | 7 (21.9) | 0.822† |
| Ascites preoperative n (%) | 7 (4.5) | 2 (1.6) | 5 (15.6) | 0.005† |
| Liver cirrhosis n (%) | 7 (4.5) | 5 (4.1) | 2 (6.3) | 0.603 |
| MELD score no of missing values: 5 | 7 (7, 9) | 7 (7, 8) | 8 (7, 9) | 0.104* |
| Child-Pugh Score n (%) | <0.001† | |||
| No of missing values: 4 | ||||
| Child A | 145 (94.2) | 119 (97.5) | 26 (81.3) | |
| Child B | 5 (3.2) | 1 (0.8) | 4 (12.5) | |
| Creatinine, mg/dL no of missing values: 1 | 0.77 (0.64, 0.91) | 0.79 (0.66, 0.93) | 0.68 (0.57, 0.78) | 0.002* |
| Bilirubin, mg/dL no of missing values: 1 | 0.6 (0.4, 0.82) | 0.6 (0.4, 0.8) | 0.62 (0.41, 1.21) | 0.326* |
| AST, U/L no of missing values: 1 | 36 (26, 47.5) | 33 (25.5, 44.5) | 47 (31.25, 74.75) | 0.015* |
| ALT, U/L no of missing values: 1 | 33 (23, 52.5) | 30 (22.5, 47.5) | 47.5 (24.5, 75.5) | 0.019* |
| ALP, U/L no of missing values: 17 | 103 (73, 163.5) | 100 (72, 147.5) | 128 (97.5, 281.8) | 0.027* |
| gGT, U/L no of missing values: 3 | 96 (45, 201) | 84 (37, 172) | 137.5 (79.8, 248) | 0.004* |
| INR no of missing values: 3 | 1.02 (0.96, 1.07) | 1.01 (0.96, 1.06) | 1.04 (0.99, 1.16) | 0.056* |
| CRP, mg/L no of missing values: 1 | 4.8 (1.7, 15.5) | 4.5 (1.5, 14.1) | 9.15 (2.32, 20.5) | 0.081* |
| TACE n (%) | 2 (1.3) | 2 (1.6) | 0 (0.0) | 0.466† |
| RFA n (%) | 1 (0.6) | 1 (0.8) | 0 (0.0) | 0.607† |
| PVE n (%) | 29 (18.8) | 21 (17.2) | 8 (25.0) | 0.316† |
| Diagnosis | 0.807† | |||
| Malignant n (%) | ||||
| HCC | 25 (16.2) | 19 (15.5) | 6 (18.8) | |
| CLC | 50 (32.5) | 37 (30.3) | 13 (40.6) | |
| Gall bladder cancer | 1 (0.6) | 0 (0) | 1 (3.1) | |
| Metastases n (%) | ||||
| CRC | 38 (24.7) | 31 (25.4) | 7 (21.9) | |
| NET | 5 (3.2) | 5 (4.1) | 0 (0) | |
| GIST | 2 (1.3) | 2 (1.6) | 0 (0) | |
| RCC | 2 (1.3) | 2 (1.6) | 0 (0) | |
| Ovarian cancer | 2 (1.3) | 1 (0.8) | 1 (3.1) | |
| Breast cancer | 1 (0.6) | 1 (0.8) | 0 (0) | |
| Choroidal melanoma | 1 (0.6) | 1 (0.8) | 0 (0) | |
| Thymus cancer | 1 (0.6) | 1 (0.8) | 0 (0) | |
| Leiomyosarcoma | 1 (0.6) | 1 (0.8) | 0 (0) | |
| Benign n (%) | ||||
| Hepatocellular adenoma | 7 (4.5) | 6 (4.9) | 1 (3.1) | |
| Angioleiomyolipoma | 1 (0.6) | 1 (0.8) | 0 (0) | |
| FNH | 3 (1.9) | 3 (2.5) | 0 (0) | |
| Mucinous/biliary cystadenoma | 2 (1.3) | 2 (1.6) | 0 (0) | |
| Infections n (%) | 3 (1.9) | 2 (1.6) | 1 (3.1) | |
| Miscellaneous n (%) | ||||
| Unknown origin | 1 (0.6) | 1 (0.8) | 0 (0) | |
| Other | 5 (3.2) | 3 (2.5) | 2 (6.3) | |
| Living donor hepatectomy n (%) | 3 (1.9) | 3 (2.5) | 0 (0) | |
| Type of liver resection | 0.237‡ | |||
| Two or fewer segments, miscellaneous n (%) | 32 (20.8) | 25 (20.5) | 7 (21.9) | |
| Hemihepatectomy n (%) | 61 (39.6) | 54 (44.3) | 7 (21.9) | |
| Extended hemihepatectomy n (%) | 61 (39.6) | 43 (35.2) | 18 (56.3) | |
| Reresection n (%) | 7 (4.5) | 6 (4.9) | 1 (3.1) | 0.665† |
| Laparoscopic surgery n (%) | 14 (9.1) | 12 (9.8) | 2 (6.3) | 0.530† |
| Duration of surgery, min | 271 (210, 341) | 255 (209, 331) | 301 (248, 378) | 0.039* |
| Haematocrit, % | 36.0 (34.0, 38.0) | 36.0 (34.0, 39.0) | 34.0 (31.3, 37.0) | 0.018* |
Data are shown as medians with quartiles (25%, 75%) or as frequencies n (%).
*P values were calculated using the Wilcoxon-Mann-Whitney U test.
†χ2 test.
‡Linear-by-linear association.
ALP, alkaline phosphatase; ALT, alanine aminotransferase; ASA, American Society of Anesthesiologists; AST, aspartate aminotransferase; AUDIT, Alcohol Use Disorders Identification Test; BMI, body mass index; CLC, Cholangiocellular carcinoma; CRC, Colorectal cancer; CRP, C reactive protein; FNH, focal nodular hyperplasia; gGT, gamma-glutamyl transferase; GIST, gastrointestinal stromal tumour; HCC, hepatocellular carcinoma; INR, international normalised ratio; MELD, Model of End-Stage Liver Disease; NET, neuroendocrine tumour; NYHA, New York Heart Association; PVE, Portal vein embolization; RCC, renal cell cancer; RFA, radiofrequency ablation; TACE, transarterial chemoembolisation.
The length of ICU stay and total hospital LOS was longer in patients with hypoalbuminaemia than in those without hypoalbuminaemia (table 2). Postoperative complications were more frequent and of higher severity in the hypoalbuminaemia group (table 2).
Table 2.
Postoperative patient characteristics
| Full sample N=154 | No hypoalbuminaemia n=122 | Hypoalbuminaemia n=32 | P value | |
| Postoperative ascites n (%) | 23 (14.9) | 16 (13.1) | 7 (21.9) | 0.264† |
| Postoperative administration of albumin n (%) | 47 (30.5) | 33 (27.0) | 14 (43.8) | 0.085† |
| Albumin dosage g/L N=47 |
270 (80, 665) | 250 (85, 655) Missing: 1 |
380 (75, 705) | 0.684* |
| No of postoperative complications | 3 (2, 4) | 3 (1, 4) | 4 (2, 7) | 0.004* |
| Highest grade of postoperative complications according to Clavien | 0.012‡ | |||
| No complication | 13 (8.4) | 11 (9.0) | 2 (6.3) | |
| Grade I, n (%) | 49 (31.8) | 44 (36.1) | 5 (15.6) | |
| Grade II, n (%) | 55 (35.7) | 43 (35.2) | 12 (37.5) | |
| Grade III, n (%) | 31 (20.1) | 20 (16.4) | 11 (34.4) | |
| Grade IV, n (%) | 6 (3.9) | 4 (3.3) | 2 (6.3) | |
| Grade V, n (%) | 0 (0.0) | 0 (0.0) | 0 (0.0) | |
| ICU LOS, hours | 23 (20, 64) | 22 (20, 44) | 32 (21, 117) | 0.005* |
| Total hospital LOS, days | 12.0 (8.8, 23.0) | 11.0 (8.0, 20.0) | 16.5 (11.5, 29.5) | 0.011* |
Data are shown as medians with quartiles (25%, 75%) or as frequencies n (%).
*P values were calculated using the Wilcoxon-Mann-Whitney U test.
†χ2 test.
‡Linear-by-linear association.
ICU, intensive care unit; LOS, length of stay.
Hypoalbuminaemia is associated with severe postoperative complications
Binary logistic regression analysis revealed significant associations between preoperative serum albumin and severe postoperative complications (Clavien grade ≥III) (OR 3.051 (95% CI 1.197 to 7.775); p=0.019). These results were independent of age, sex, randomisation, ASA status, preoperative surgical diagnosis and Child-Pugh class (table 3). In a second step, we added duration and type of surgery as independent predictors to the model to control for intraoperative confounders. Hypoalbuminaemia remained significantly associated with the occurrence of postoperative complications ≥Clavien grade III (OR 3.148 (95% CI 1.194 to 8.303); p=0.020) (online supplemental table S1).
Table 3.
Logistic regression analysis with postoperative complications (Clavien grade ≥ III) as dependent variable
| Postoperative complications (Clavien grade ≥III) | OR | 95% CI | P value |
| Hypoalbuminaemia | 3.051 | 1.197 to 7.775 | 0.019 |
| Sex (female #) | 0.826 | 0.345 to 1.975 | 0.667 |
| Age | 1.003 | 0.962 to 1.003 | 0.905 |
| Randomisation (placebo vs physostigmine) |
1.209 | 0.541 to 2.701 | 0.643 |
| Diagnosis | |||
| Benign # | |||
| Malignant | 1.783 | 0.269 to 11.822 | 0.549 |
| Metastases | 0.867 | 0.122 to 6.169 | 0.887 |
| Miscellaneous | 2.944 | 0.358 to 24.224 | 0.315 |
| Child Pugh Class Child A # |
0.331 | 0.030 to 3.652 | 0.367 |
| ASA | |||
| ASA I # | |||
| ASA II | 1.925 | 0.299 to 12.412 | 0.491 |
| ASA III | 1.043 | 0.109 to 9.950 | 0.971 |
#, reference; ASA, American Society of Anesthesiologists.
bmjopen-2022-068405supp001.pdf (77.6KB, pdf)
Hypoalbuminaemia is associated with prolonged LOS
Hypoalbuminaemia was associated with both longer ICU and total hospital LOS independent of age, sex, randomisation, ASA status, preoperative surgical diagnosis and Child-Pugh class (OR 2.573 (95% CI 1.015 to 6.524); p=0.047 and OR 1.296 (95% CI 0.254 to 3.009); p=0.012, respectively) (tables 4 and 5). Of note, when additionally adjusting for type and length of surgery, the association between preoperative hypoalbuminaemia and ICU LOS did no longer remain statistically significant (OR 2.489 (95% CI 0.948 to 6.537); p=0.064) (online supplemental table S2). However, as for hospital LOS, results remained significant after adjusting for the type and duration of surgery (OR 1.324 (95% CI 0.080 to 3.407); p=0.032) (online supplemental table S3).
Table 4.
Logistic regression analysis with length of ICU stay as dependent variable
| Length of ICU stay | OR | 95% CI | P value |
| Hypoalbuminaemia | 2.573 | 1.015 to 6.524 | 0.047 |
| Sex (female*) | 0.897 | 0.408 to 1.970 | 0.786 |
| Age | 1.020 | 0.983 to 1.058 | 0.295 |
| Randomisation (placebo vs physostigmine) |
1.275 | 0.628 to 2.586 | 0.501 |
| Diagnosis | |||
| Benign* | |||
| Malignant | 3.988 | 0.741 to 20.876 | 0.108 |
| Metastases | 1.178 | 0.214 to 6.498 | 0.851 |
| Miscellaneous | 1.384 | 0.204 to 9.377 | 0.739 |
| Child Pugh Class Child A* |
1.245 | 0.117 to 13.284 | 0.856 |
| ASA | |||
| ASA I* | |||
| ASA II | 0.607 | 0.127 to 2.901 | 0.532 |
| ASA III | 0.425 | 0.066 to 2.724 | 0.367 |
*Reference.
ASA, American Society of Anesthesiologists; ICU, intensive care unit.
Table 5.
Logistic regression analysis with hospital length of stay as dependent variable
| Hospital length of stay | OR | 95% CI | P value |
| Hypoalbuminaemia | 1.296 | 0.254 to 3.009 | 0.012 |
| Sex (female*) | –0.528 | –1.795 to 0.431 | 0.281 |
| Age | 0.011 | –0.038 to 0.057 | 0.612 |
| Randomisation (placebo vs physostigmine) |
–0.522 | –1.559 to 0.202 | 0.199 |
| Diagnosis | |||
| Benign* | |||
| Malignant | 21.984 | 20.708 to 24.784 | 0.002 |
| Metastases | 20.560 | 19.148 to 22.556 | 0.002 |
| Miscellaneous | 20.784 | 18.316 to 22.886 | 0.001 |
| Child Pugh Category Child A* |
0.130 | –3.337 to 37.306 | 0.595 |
| ASA | |||
| ASA I* | |||
| ASA II | 0.573 | –2.102 to 21.713 | 0.556 |
| ASA III | 0.236 | –3.012 to 21.544 | 0.768 |
*Reference.
ASA, American Society of Anesthesiologists.
Preoperative hypoalbuminaemia is not associated with 1-year survival after surgery
Overall, 24 (15.6%) patients died within 1 year. Kaplan-Meier analysis revealed no difference in 1-year survival between patients with normal preoperative albumin and those with hypoalbuminaemia (17 (13.9%) vs 7 (21.9%); p=0.256). Cox regression analysis revealed no association between preoperative hypoalbuminaemia and 1-year survival while adjusting for age, sex, ASA, Charlson Comorbidity Index (CCI) and Child-Pugh class (HR 2.175 (95% CI 0.812 to 5.825); p=0.122) (online supplemental table S4).
Discussion
Our investigation of patients undergoing partial hepatectomy revealed that preoperative hypoalbuminaemia is associated with severe postoperative complications as well as total hospital and ICU LOS. However, no such association was observed for 1-year survival. These findings were independent of age, sex, randomisation, ASA status, preoperative diagnosis and Child-Pugh class. Importantly, when adjusting additionally for surgery type and duration, the association between preoperative hypoalbuminaemia and ICU LOS failed to reach statistical significance. In all other analyses, preoperative hypoalbuminaemia remained a significant predictor for short-term outcomes. To the best of our knowledge, this is the first investigation including a mixed liver surgical cohort demonstrating that preoperative hypoalbuminaemia is associated with adverse outcomes independent of the underlying condition.
Although it has been previously demonstrated that patients undergoing elective orthopaedic and neurosurgery with preoperative hypoalbuminaemia are linked to an increased risk of postoperative complications and longer hospital and ICU stays,19 20 data on the impact of hypoalbuminaemia on the postoperative outcome of liver surgery are limited. Wang et al24 investigated patients undergoing liver resection for HCC. The authors found the albumin-bilirubin ratio to be more predictive for posthepatectomy liver failure than the Child-Pugh score. These findings underscore the value of serum albumin for preoperative risk stratification of patients undergoing partial hepatectomy. Cumulative long-term survival has been shown to be reduced in patients with HCC and concomitant hypoalbuminaemia.22 23 30 However, we failed to confirm this association. The inclusion of patients with both benign and malignant disorders might explain our findings, even though the number of benign diagnoses was equally distributed between patients with and without hypoalbuminaemia. However, we believe that including patients with malignancies only, as in previous studies,22–24 30 ignores a significant patient population undergoing liver surgery.
In addition, the results reported by others were mostly unadjusted for preexisting medical conditions, whereas we included ASA physical status and CCI as surrogates for preoperative comorbidities in the multivariable analysis. Given the available evidence, it is widely accepted that the presence of hypoalbuminaemia before surgery is associated with increased mortality in both patients with cancer and patients undergoing cardiac, vascular and neurosurgery.13 15 18
The mechanisms explaining the impact of hypoalbuminaemia on adverse postoperative outcomes include proinflammation and loss of control over cell proliferation.31 Impaired collagen synthesis32 and impaired macrophage activation lead to an insufficient immune response,9 thereby contributing to surgical site infections. In addition, loss of albumin is directly related to a shortage of buffer, resulting in acidosis33 and an increased risk of postoperative complications.
Given that more than 80% of our study population had an underlying malign condition, preoperative chronic inflammation was likely present in many patients.34 In these cases, cytokine release, particularly interleukin-6 (IL-6), affects capillary integrity, thereby facilitating albumin extravasation and subsequently reducing the serum albumin concentration.35 36 Moreover, IL-6 directly inhibits hepatic albumin synthesis.37 38
Preoperative ascites was more common among hypoalbuminaemic patients than in non-hypoalbuminaemic patients (15.6% vs 1.6%; p=0.005), while the occurrence of postoperative ascites showed no significant difference between patient types (21.9% vs 13.1%; p=0.085). Thus, patients’ ability to cope with postoperative stress may not be the decreased hepatic function in response to liver resection but rather the preoperatively reduced liver synthetic capacity.
Few studies have assessed the impact of preoperative hypoalbuminaemia on the postoperative outcome of liver surgery, most of which are retrospective in design. Moreover, we systematically classified postoperative complications according to Clavien,29 while others have only assessed infectious complications.19 20 As albumin has anti-inflammatory properties, it might be hypothesised that preoperative correction of hypoalbuminaemia improves postoperative outcome.39 40 Thus, the impact of preoperative albumin substitution on inflammatory markers and outcomes such as postoperative infections and metastases needs further investigation.
Our study is limited because the low level of albumin could also have been due to body water imbalances leading to pseudohypoalbuminaemia. However, we tried to rule out such bias by excluding all patients with a haematocrit of more than 10% below the lower reference limit. By adjusting for ASA status and CCI we accounted for preexisting conditions although we did not adjust for liver cirrhosis. Yet, we believe that with confirmed operability the number of patients with severe cirrhosis is negligible. A considerable number of exclusion criteria in the original trial might limit the generalisability of our results. We investigated a selected patient cohort who had highly specialised surgery. Originally, this cohort was selected as these patients were at high risk to develop postoperative cognitive impairments, that is, postoperative delirium and neurocognitive disorders due to extended liver surgery. Yet, we believe that the findings of our subgroup analysis are applicable to patients undergoing extended liver resection as surgical trauma and anaesthetic regimen were comparable in our cohort and findings with regard to preoperative hypoalbuminaemia are unaffected by the trial’s exclusion criteria. Generalisability could be furthermore limited by the rather small number of patients presenting with preoperative hypoalbuminaemia (n=32). Therefore, further investigation of this particular patient group is warranted.
Hypoalbuminaemia is interpreted cautiously in liver surgery, as it reflects compromised liver function and thereby disease severity, prompting worse postoperative outcomes. However, the multivariable analysis confirmed that the association of preoperative hypoalbuminaemia and postoperative outcome was independent of the preoperative diagnosis and surgery type.
Our findings underscore the predictive value of serum albumin, as its preoperative assessment may help in the risk stratification of patients undergoing liver resection. Whether the correction of preoperative hypoalbuminaemia improves postoperative outcome remains to be investigated in future studies.
Conclusions
Preoperative hypoalbuminaemia has been identified as a risk factor for a worse short-term outcome after partial hepatectomy independent of age, sex, randomisation, ASA status, preoperative diagnosis or type and duration of surgery.
Supplementary Material
Footnotes
Contributors: Conceived and designed the study: ADK, CS and CL. Obtained the data: ADK, CvH, MM, LAS, NW and CL. Analysed the data: ADK, ES, MJ, GL and CL. Wrote the manuscript: ADK, ES, MJ, CL and GL. Commented on the manuscript: all authors. CL accepts full responsibility for the work and the conduct of the study, had access to the data, and controlled the decision to publish.
Funding: Gunnar Lachmann is participant in the BIH Charité Clinician Scientist Program funded by the Charité – Universitätsmedizin Berlin, and the BIH at Charité. Cornelia Lachmann is participant in the Berlin Institute of Health (BIH) Charité Digital Clinician Scientist Program funded by the Charité – Universitätsmedizin Berlin, and the BIH at Charité. The investigator-initiated trial PHYDELIO was funded by Köhler Pharma, Alsbach Hähnlein, Germany.
Competing interests: The authors declare no competing interests.
Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review: Not commissioned; externally peer reviewed.
Supplemental material: This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
Data availability statement
Data are available on reasonable request. No data are available.
Ethics statements
Patient consent for publication
Not applicable.
Ethics approval
This study involves human participants this study was approved by the 'Landesamt für Gesundheit und Soziales Berlin' (Berlin, Germany) ethics committee on 15 January 2009 (Ethics approval ZS EK 11618/08). Participants gave informed consent to participate in the study before taking part.
References
- 1.Horwich TB, Kalantar-Zadeh K, MacLellan RW, et al. Albumin levels predict survival in patients with systolic heart failure. Am Heart J 2008;155:883–9. 10.1016/j.ahj.2007.11.043 [DOI] [PubMed] [Google Scholar]
- 2.Don BR, Kaysen G. Serum albumin: relationship to inflammation and nutrition. Semin Dial 2004;17:432–7. 10.1111/j.0894-0959.2004.17603.x [DOI] [PubMed] [Google Scholar]
- 3.Fuhrman MP, Charney P, Mueller CM. Hepatic proteins and nutrition assessment. J Am Diet Assoc 2004;104:1258–64. 10.1016/j.jada.2004.05.213 [DOI] [PubMed] [Google Scholar]
- 4.Gabay C, Kushner I. Acute-Phase proteins and other systemic responses to inflammation. N Engl J Med 1999;340:448–54. 10.1056/NEJM199902113400607 [DOI] [PubMed] [Google Scholar]
- 5.De Feo P, Lucidi P. Liver protein synthesis in physiology and in disease states. Curr Opin Clin Nutr Metab Care 2002;5:47–50. 10.1097/00075197-200201000-00009 [DOI] [PubMed] [Google Scholar]
- 6.Kaysen GA, Dubin JA, Müller H-G, et al. Inflammation and reduced albumin synthesis associated with stable decline in serum albumin in hemodialysis patients. Kidney Int 2004;65:1408–15. 10.1111/j.1523-1755.2004.00520.x [DOI] [PubMed] [Google Scholar]
- 7.Zhang WJ, Frei B. Albumin selectively inhibits TNF alpha-induced expression of vascular cell adhesion molecule-1 in human aortic endothelial cells. Cardiovasc Res 2002;55:820–9. 10.1016/s0008-6363(02)00492-3 [DOI] [PubMed] [Google Scholar]
- 8.Sun HY, Singh N, Cacciarelli TV, et al. Dysregulated expression of T-helper cell responses and susceptibility to infections in high-risk liver transplant recipients. Transpl Immunol 2008;20:68–72. 10.1016/j.trim.2008.08.001 [DOI] [PubMed] [Google Scholar]
- 9.Reynolds JV, Redmond HP, Ueno N, et al. Impairment of macrophage activation and granuloma formation by protein deprivation in mice. Cell Immunol 1992;139:493–504. 10.1016/0008-8749(92)90088-7 [DOI] [PubMed] [Google Scholar]
- 10.Rivadeneira DE, Grobmyer SR, Naama HA, et al. Malnutrition-induced macrophage apoptosis. Surgery 2001;129:617–25. 10.1067/msy.2001.112963 [DOI] [PubMed] [Google Scholar]
- 11.Veenhof AAFA, Sietses C, von Blomberg BME, et al. The surgical stress response and postoperative immune function after laparoscopic or conventional total mesorectal excision in rectal cancer: a randomized trial. Int J Colorectal Dis 2011;26:53–9. 10.1007/s00384-010-1056-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Yamamoto M, Shimura T, Kano S, et al. Prognostic value of hypoalbuminemia after transcatheter aortic valve implantation (from the Japanese multicenter OCEAN-TAVI registry). The American Journal of Cardiology 2017;119:770–7. 10.1016/j.amjcard.2016.11.019 [DOI] [PubMed] [Google Scholar]
- 13.Kato TS, Cheema FH, Yang J, et al. Preoperative serum albumin levels predict 1-year postoperative survival of patients undergoing heart transplantation. Circ: Heart Failure 2013;6:785–91. 10.1161/CIRCHEARTFAILURE.111.000358 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kato TS, Kitada S, Yang J, et al. Relation of preoperative serum albumin levels to survival in patients undergoing left ventricular assist device implantation. The American Journal of Cardiology 2013;112:1484–8. 10.1016/j.amjcard.2013.06.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Inagaki E, Farber A, Eslami MH, et al. Preoperative hypoalbuminemia is associated with poor clinical outcomes after open and endovascular abdominal aortic aneurysm repair. Journal of Vascular Surgery 2017;66:53–63. 10.1016/j.jvs.2016.10.110 [DOI] [PubMed] [Google Scholar]
- 16.Peacock MR, Farber A, Eslami MH, et al. Hypoalbuminemia predicts perioperative morbidity and mortality after infrainguinal lower extremity bypass for critical limb ischemia. Annals of Vascular Surgery 2017;41:169–175. 10.1016/j.avsg.2016.08.043 [DOI] [PubMed] [Google Scholar]
- 17.Rady MY, Ryan T, Starr NJ. Clinical characteristics of preoperative hypoalbuminemia predict outcome of cardiovascular surgery. JPEN J Parenter Enteral Nutr 1997;21:81–90. 10.1177/014860719702100281 [DOI] [PubMed] [Google Scholar]
- 18.Kim K, Bang JY, Kim SO, et al. n.d. Association of preoperative hypoalbuminemia with postoperative acute kidney injury in patients undergoing brain tumor surgery: a retrospective study. J Neurosurg;128:1115–22. 10.3171/2016.11.JNS162237 [DOI] [PubMed] [Google Scholar]
- 19.Adogwa O, Martin JR, Huang K, et al. Preoperative serum albumin level as a predictor of postoperative complication after spine fusion. Spine 2014;39:1513–9. 10.1097/BRS.0000000000000450 [DOI] [PubMed] [Google Scholar]
- 20.Bohl DD, Shen MR, Kayupov E, et al. Hypoalbuminemia independently predicts surgical site infection, pneumonia, length of stay, and readmission after total joint arthroplasty. The Journal of Arthroplasty 2016;31:15–21. 10.1016/j.arth.2015.08.028 [DOI] [PubMed] [Google Scholar]
- 21.Pimlott BJ, Jones CA, Beaupre LA, et al. Prognostic impact of pre-operative albumin on short-term mortality and complications in patients with hip fracture. Arch Gerontol Geriatr 2011;53:90–4. 10.1016/j.archger.2010.06.018 [DOI] [PubMed] [Google Scholar]
- 22.Chen M-F, Tsai H-P, Jeng L-B, et al. Prognostic factors after resection for hepatocellular carcinoma in noncirrhotic livers: univariate and multivariate analysis. World Journal of Surgery 2003;27:443–7. 10.1007/s00268-002-6708-7 [DOI] [PubMed] [Google Scholar]
- 23.Lee C-W, Tsai H-I, Sung C-M, et al. Risk factors for early mortality after hepatectomy for hepatocellular carcinoma. Medicine (Baltimore) 2016;95:e5028. 10.1097/MD.0000000000005028 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Wang Y-Y, Zhong J-H, Su Z-Y, et al. Albumin-Bilirubin versus Child-Pugh score as a predictor of outcome after liver resection for hepatocellular carcinoma. Br J Surg 2016;103:725–34. 10.1002/bjs.10095 [DOI] [PubMed] [Google Scholar]
- 25.Network e . The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. n.d. Available: https://www.equator-network.org/reporting-guidelines/strobe/ [DOI] [PMC free article] [PubMed]
- 26.Spies CD, Knaak C, Mertens M, et al. Physostigmine for prevention of postoperative delirium and long-term cognitive dysfunction in liver surgery. Eur J Anaesthesiol 2021;38:943–56. 10.1097/EJA.0000000000001456 [DOI] [PubMed] [Google Scholar]
- 27.Spies C, Kastrup M, Kerner T, et al. SOPs in der Anästhesiologie und Schmerztherapie. Stuttgart: Georg Thieme Verlag KG, 2013. [Google Scholar]
- 28.Vivantes LBKvCu . Hämatokrit: Hämatologie-allgemeine Hämatologie-kleines blutbild. Berlin. n.d. Available: https://www.laborberlin.com/unsere-leistungen/leistungsverzeichnis-detail/?id=438 [Google Scholar]
- 29.Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004;240:205–13. 10.1097/01.sla.0000133083.54934.ae [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Deng Y, Pang Q, Miao R-C, et al. Prognostic significance of pretreatment albumin/globulin ratio in patients with hepatocellular carcinoma. Onco Targets Ther 2016;9:5317–28. 10.2147/OTT.S109736 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Laursen I, Briand P, Lykkesfeldt AE. Serum albumin as a modulator on growth of the human breast cancer cell line, MCF-7. Anticancer Res 1990;10:343–51. [PubMed] [Google Scholar]
- 32.Kim J, Shim S-H, Oh I-K, et al. Preoperative hypoalbuminemia is a risk factor for 30-day morbidity after gynecological malignancy surgery. Obstet Gynecol Sci 2015;58:359. 10.5468/ogs.2015.58.5.359 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Seaton K. Albumin concentration controls cancer. J Natl Med Assoc 2001;93:490–3. [PMC free article] [PubMed] [Google Scholar]
- 34.McMillan DC, Watson WS, O’Gorman P, et al. Albumin concentrations are primarily determined by the body cell mass and the systemic inflammatory response in cancer patients with weight loss. Nutr Cancer 2001;39:210–3. 10.1207/S15327914nc392_8 [DOI] [PubMed] [Google Scholar]
- 35.Gupta D, Lis CG. Pretreatment serum albumin as a predictor of cancer survival: a systematic review of the epidemiological literature. Nutr J 2010;9:69. 10.1186/1475-2891-9-69 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Nisar PJ, Appau KA, Remzi FH, et al. Preoperative hypoalbuminemia is associated with adverse outcomes after ileoanal pouch surgery. Inflamm Bowel Dis 2012;18:1034–41. 10.1002/ibd.21842 [DOI] [PubMed] [Google Scholar]
- 37.Heinrich PC, Castell JV, Andus T. Interleukin-6 and the acute phase response. Biochem J 1990;265:621–36. 10.1042/bj2650621 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Steel DM, Whitehead AS. The major acute phase reactants: C-reactive protein, serum amyloid P component and serum amyloid a protein. Immunol Today 1994;15:81–8. 10.1016/0167-5699(94)90138-4 [DOI] [PubMed] [Google Scholar]
- 39.Arroyo V, García-Martinez R, Salvatella X. Human serum albumin, systemic inflammation, and cirrhosis. Journal of Hepatology 2014;61:396–407. 10.1016/j.jhep.2014.04.012 [DOI] [PubMed] [Google Scholar]
- 40.Fernández J, Clària J, Amorós A, et al. Effects of albumin treatment on systemic and portal hemodynamics and systemic inflammation in patients with decompensated cirrhosis. Gastroenterology 2019;157:149–62. 10.1053/j.gastro.2019.03.021 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
bmjopen-2022-068405supp001.pdf (77.6KB, pdf)
Data Availability Statement
Data are available on reasonable request. No data are available.