Impact of Frailty on Short-Term Outcomes After Laparoscopic and Open Hepatectomy

Background Although laparoscopic hepatectomy (LH) is associated with improved short-term outcomes compared to open hepatectomy (OH), it is unknown whether frail patients also benefit from LH. The aim of this study was to evaluate the impact of frailty on post-operative outcomes after LH and OH. Patients and methods Consecutive patients who underwent LH and OH between January 2011 and December 2018 were identified from a prospective database. Frailty was assessed using the modified Frailty Index (mFI), with patients scoring mFI ≥ 1 deemed to be frail. Results Of 1826 patients, 34.7% (N = 634) were frail and 18.6% (N = 340) were elderly (≥ 75 years). Frail patients had significantly higher 90-day mortality (6.6% vs. 2.9%, p < 0.001) and post-operative complications (36.3% vs. 26.1%, p < 0.001) than those who were not frail, effects that were independent of patient age on multivariate analysis. For those undergoing minor resections, the benefits of LH vs. OH were similar for frail and non-frail patients. Length of hospital stay was 53% longer in OH (vs. LH) in frail patients, compared to 58% longer in the subgroup of non-frail patients. Conclusions Frailty is independently associated with inferior post-operative outcomes in patients undergoing hepatectomy. However, the benefits of laparoscopic (compared to open) hepatectomy are similar for frail and non-frail patients. Frailty should not be a contraindication to laparoscopic minor hepatectomy in carefully selected patients.


Introduction
Hepatectomy is a potentially curative treatment for patients with primary and secondary hepatic malignancies, and a laparoscopic approach, when feasible, is associated with less morbidity, lower mortality and faster recovery than open surgery [1,2]. In an increasingly ageing population [3], there is a growing number of frail and elderly patients with comorbidity who are being considered for major surgery, including hepatectomy [4]. Although the risks of surgery are known to increase with advancing age, the short-term advantages of laparoscopic hepatectomy (LH) compared to open hepatectomy (OH) appear to be retained, at least in selected patients undergoing minor hepatectomy [5,6]. However, the benefits of LH appear to diminish with increasing age amongst elderly cohorts [7], and it is not known whether similar benefits of LH are observed in frail patients.
Frailty is a condition that is characterised by reduced physiological reserve and is associated with an increased risk of post-operative complications, prolonged hospitalization, increased readmission rates and loss of independence following general surgical procedures [8]. Frailty is also common in patients undergoing hepatectomy, with a reported incidence of 14-29% in recently published studies [9,10], although there is currently limited data to support an association with worse post-operative outcomes [10]. Available data on the impact of frailty in minimally invasive surgery is also limited and conflicting. In two recent studies of the ACS-NSQIP database, Kothari et al. demonstrated that the benefits of laparoscopic colectomy over open colectomy were preserved in frail patients [11], whilst Lo et al. reported possible worse outcomes after robotic compared to open colectomy [12]. The effects of frailty on the short-term outcomes of LH are unknown, and it is unclear whether frail patients should preferentially undergo LH or OH.
Various clinical scores have been developed and validated to measure frailty in surgical patients, and typically include an assessment of functional status combined with the presence and severity of comorbidity, such as diabetes mellitus, hypertension and cardiovascular disease [13]. The modified Frailty Index (mFI) [14] consists of 11 variables that were adapted from the original 70-item Frailty Index [15] and has been shown to be significantly associated with poor surgical outcomes, including after hepatectomy [8].
The primary aim of this study was to evaluate the effects of age and frailty on short-term post-operative outcomes after hepatectomy. The secondary aim was to evaluate the relationship between age, frailty and post-operative outcomes in patients undergoing laparoscopic and open hepatectomy. Consecutive patients who underwent laparoscopic or open  hepatectomy at a single centre between 2011 and 2018 were identified from a prospectively maintained database. Repeat hepatectomy (N = 152) and emergency hepatectomy for trauma (N = 4) patients were excluded. The type of hepatectomy was defined as minor, major or extra major according to the Tokyo 2020 terminology of liver anatomy and resections which is an update of the Brisbane 2000 system [16]. Frailty was defined as a modified Frailty Index (mFI) C 1 and was calculated for all patients. Patients aged 75 years and over were considered elderly, and the effects of both frailty and age on hepatectomy outcomes were compared. Data were collected regarding comorbidity, indications for surgery, post-operative complications, length of hospital (LOS) and 90-day mortality. Post-operative complications were graded according to the Clavien-Dindo classification [17].

Statistical analysis
Continuous variables were reported as median (interquartile range; IQR), and associations with age and mFI were assessed using Mann-Whitney U tests. Ordinal variables were analysed using the same approach, whilst nominal variables were assessed using Fisher's exact tests. To assess the interplay between age and frailty with respect to dichotomous patient outcomes, binary logistic regression models were produced with age, frailty and the age*frailty interaction as independent variables. These models were then evaluated to produce odds ratios for frail vs. non-frail patients within each age subgroup, with the p-value of the interaction term representing the comparison between these two odds ratios. Length of stay was then analysed using a similar approach, but using an ANOVA model. Lengths of stay were log 10 -transformed, prior to this analysis, in order to normalise the distribution; hence were summarised using geometric means, with differences between groups reported as percentages. Multivariable models were then produced, to assess whether age and frailty were independent predictors of the primary outcomes. This used binary logistic regression models for 90-day mortality and complication rates, with a general linear model used for length of stay, which was log 10 -transformed for analysis. Age and frailty were entered into the models as continuous covariates, and a backwards stepwise approach was used to select other potentially confounding factors for inclusion in the models. Analyses were then performed to compare the effect of operative approach (open vs. laparoscopic) between subgroups of age and frailty, which used a similar approach to that previously described. All analyses were performed using IBM SPSS 24 (IBM Corp. Armonk, NY), with p \ 0.05 deemed to be indicative of statistical significance throughout.

Age and frailty of the cohort
The majority of patients scored 0 on the mFI (65.3%), with 25.9%, 8.5%, and 0.3% scoring 1, 2 and 3 points, respectively. The hypertension and diabetes mellitus components were the main contributors to the mFI scores, being present in 26.1% and 13.3% of patients, respectively. Only 0.3% (N = 5) of patients were non-independent with their activities of daily living. A significant correlation between age and mFI was observed (rho: 0.304, p \ 0.001), with the median age increasing from 62.1 (IQR: 52.2-70.9) to 71.7 (64.4-76.7) years, for those with an mFI of 0 vs. 2-3. This was largely as a result of the comorbidity components of the mFI, with rates of hypertension, diabetes mellitus and COPD/pneumonia all increasing significantly with age (all p \ 0.001).

Characteristics and outcomes by age and frailty
For the initial analyses, both age and mFI were dichotomised, with age C 75 years classified as ''elderly'' (18.6%; N = 340), and mFI C 1 classified as ''frail'' (34.7%; N = 634). Of the demographic factors considered (  Univariable analysis of post-operative outcomes found overall complication rates to be significantly higher in both elderly (34.1% vs. 28.6%, p = 0.048) and frail (36.3% vs. 26.1%, p \ 0.001) patients (Table 2). Further assessment of complication types found that this was largely a result of higher rates of medical complications (p \ 0.001 for both), with surgical complication rates not found to be significantly higher in either elderly (p = 0.270) or frail (p = 0.061) patients. Elderly patients also had significantly longer lengths of stay (median: 7 vs. 6 days, p \ 0.001), with no such difference observed for frail patients (median: 6 vs. 6 days, p = 0.060). Mortality rates were also significantly higher in the elderly and the frail, with 90-day mortality of 6.8% vs. 3.6% (p = 0.015) for elderly vs. nonelderly, and 6.6% vs. 2.9% (p \ 0.001) for frail vs. nonfrail.

Interplay between age and frailty
Comparisons of outcomes between frail and non-frail patients were then performed within the elderly and nonelderly patient subgroups. Analysis of mortality, complication rates and length of stay found the effect of frailty to be similar in non-elderly and elderly patients (

Associations with primary outcomes
The associations between both age and frailty, and the primary outcomes of 90-day mortality, post-operative complications and length of stay were then assessed in further detail. Age and mFI were treated as continuous covariates in these analyses and were found to be significantly associated with all three outcomes on univariable analysis (Fig. 1). On multivariable analysis, both age and mFI were found to be significant independent predictors of 90-day mortality, with odds ratios of 1.71 (95% CI: 1.32-2.22, p \ 0.001) per decade of age and 1.45 (1.04-2.03, p = 0.029) per point on the mFI (Table 4). Age and mFI were found to be significant independent predictors of post-operative complications (p \ 0.001, p = 0.002, respectively). Whilst age was found to be significantly associated with length of stay (p \ 0.001), mFI narrowly missed statistical significance in this analysis (p = 0.056).

Effect of operative approach by age and frailty in minor resections
The operative approach was found to vary significantly by the extent of the resection (p \ 0.001), with 22.3% of minor resections being laparoscopic, compared to 6.1% of major and 0.8% of extra-major resections. As such, to negate the confounding effect of the extent of resection, only the subgroup of minor resections were considered in For binary outcomes, odds ratios are from binary logistic regression models within each age subgroup, with mFI (C 1 vs. 0) as the independent variable. The p-values relate to the interaction terms of binary logistic regression models with age, frailty and the age*frailty interaction as independent variables, hence represent a comparison between the reported odds ratio for each subgroup. *Analysis of length of stay excluded those patients who died in hospital prior to discharge. For the remainder, the average lengths of stay are reported as geometric mean (95% CI). Lengths of stay were then log 10 -transformed, and analysed using an ANOVA model, parameterised as previously described, with the comparisons between groups reported as percentage differences. CI confidence interval; mFI modified frailty index the analysis (N = 970). Within this subgroup, neither complication nor mortality rates were found to differ significantly by the operative approach (  Table 4 resulted in very low statistical power, meaning that only large differences will have been detectable. As such, this analysis will be subject to an inflated false-negative rate, and the results must be interpreted with this in mind.

Discussion
Frailty is increasingly recognised as being an important determinant of post-operative outcomes in surgical patients [18], and our data have shown that frailty is an independent predictor of morbidity and mortality after hepatectomy. Over one third of patients in this large cohort were considered frail, and despite acceptable short-term outcomes, were exposed to an increased risk of medical complications, prolonged recovery and post-operative mortality compared to non-frail patients, which is consistent with published data [9,19,20]. Importantly, frailty was associated with worse outcomes for both elderly and nonelderly patients undergoing hepatectomy. Data from metaanalyses and randomized trials have shown that laparoscopic hepatectomy is associated with improved short-term outcomes compared to open hepatectomy [2,21]. Our analysis has indicated that the short-term benefits (e.g. reduced hospital stay) of laparoscopic surgery are retained in frail patients undergoing minor hepatectomy, and suggests that in carefully selected patients, frailty is not a contraindication to laparoscopic minor hepatectomy. This Analyses of 90-day mortality and complications were performed using binary logistic regression models, with coefficients representing odds ratios. Analysis of length of stay was performed using general linear models, with the log 10 [length of stay] as the dependent variable; coefficients represent percentage differences. For all outcomes, age and mFI were initially treated as continuous covariates in separate univariable models, with coefficients representing the change in the outcome per decade or per point, respectively. Multivariable models were then produced, which considered gender, BMI, ASA grade, indication for surgery, extent of resection, operative approach and duration of surgery for inclusion, with a backwards stepwise approach to variable selection. The full models are reported in Supplementary  finding is not unexpected, since the benefits of laparoscopic surgery are primarily due to reduced medical complications, which occur as a result of less post-operative pain, improved respiratory function and earlier ambulation. It is unknown whether frail patients undergoing major hepatectomy would also benefit from a laparoscopic approach. A recent patient blinded randomized trial demonstrated a faster recovery in patients undergoing laparoscopic major hepatectomy compared to open [22], but there is currently no data specifically evaluating the Only those patients with minor resections (N = 970) are included in the analysis, with analysis of length of stay additionally excluding those patients who died in hospital prior to discharge. For the analysis by age, odds ratios are from binary logistic regression models within each age subgroup, with the operative approach (open vs. laparoscopic) as the independent variable. The p-values relate to the interaction terms of binary logistic regression models with age, operative approach and the age*operative approach interaction as independent variables, hence represent a comparison between the reported odds ratio for each subgroup. The analysis was repeated similarly for mFI. *Operative duration and length of stay followed skewed distributions, hence were log 10 -transformed, and analysed using an ANOVA model, parameterised as previously described; averages are reported as geometric means (95% CI), and comparisons between groups are reported as percentage differences. **Hazard ratios were not calculable, as there were no events in one of the subgroups outcomes of laparoscopic major hepatectomy in frail patients. Laparoscopic major hepatectomy is a complex procedure associated with significantly longer operating times and longer hepatic inflow occlusion times compared to open surgery, and it is unknown whether these factors may negate the potential benefits of laparoscopic surgery in frail patients. It was not possible to evaluate the effect of laparoscopic approach in frail patients undergoing major hepatectomy in our study due to small numbers in this subgroup (Supplementary Table 4), and it is likely that a multi-centre study would be required. Detection of frailty prior to major surgery is important, since it may allow risk stratification, facilitates preoperative counselling, and guides perioperative management including choice of post-operative destination (i.e. critical care or surgical high dependency). Frail patients may also benefit from preoperative interventions to address reversible deficits (e.g. aerobic fitness and nutrition) [23], and the concept of prehabilitation is likely to become a central component of perioperative care for patients being considered for hepatectomy in the near future [24,25]. Postoperative functional recovery following hospital discharge and return to baseline function is an under-researched, patient-centred outcome that may also be influenced by frailty. A recent Japanese study of over 65-year-old patients undergoing hepatectomy found that frailty, advanced age (C 76 years) and open hepatectomy were independent risk factors for post-operative loss of independence [9].
This study has several limitations. Due to its retrospective nature, the effect of selection bias on the operative approach may have affected the results, and it was also not possible to ascertain how many patients were deemed unsuitable for hepatectomy due to severe frailty. As discussed above, the small number of laparoscopic major hepatectomy patients precluded analysis of the impact of frailty and outcome in this subgroup.
In conclusion, frailty is a common finding in patients undergoing hepatectomy, and is an independent risk factors for post-operative morbidity and mortality. The short-term benefits of laparoscopic hepatectomy appear to be preserved in frail patients. As such, frailty is not contraindicated in patients being considered for laparoscopic minor hepatectomy. Further study is needed to determine if frail patients would also benefit from laparoscopic major hepatectomy.

Declarations
Conflict of interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/.