Associations between biological age and complications after major cancer surgery

Abstract

Patients who are older are increasingly presenting for major cancer surgery. Postoperative complications are common after major surgery and are expensive for the patient (prolonged inpatient stay, increased hospital readmissions and inability to return to premorbid function) and healthcare institutions. Prehabilitation is reported to have substantial promise in reducing postoperative complications [1]. However, choosing the most appropriate patients for prehabilitation before major cancer surgery to make it cost-effective remains a challenge, and whether prehabilitation improves medium- to long-term survival remains uncertain [2, 3].

As a proof-of-concept study, we aimed to identify novel factors that are predictive of postoperative complications in patients undergoing cancer surgery so that those most at risk of poorer long-term outcomes can be selected for a randomised controlled trial assessing the possible benefits of prehabilitation. We hypothesised that increased biological age, as a surrogate for poor predictive of postoperative complications within 30 days of cancer surgery.

Approval was obtained from the Division of Surgical Oncology Research Committee at Peter MacCallum Cancer Centre and we extracted data from the electronic medical records of adult patients who underwent major cancer surgery between 2018 and 2022. Duration of surgery had to have been at least 2 h and involved one overnight stay. We used the Levine PhenoAge model to determine biological age. This model is based on nine biomarkers and has been shown to predict short- and long-term outcomes of different patient cohorts [4, 5]. Accelerated biological ageing (PhenoAgeAccel) was defined as a patient's biological age being older than their chronological age [4]. Patients who had the complete set of nine baseline blood tests used to calculate PhenoAge were included in the analysis.

Postoperative complications were defined by organ type, with severity classified using the Clavien-Dindo score [6]. These categorical outcomes were then converted into a numerical scale for comparison using the Comprehensive Complication Index [7]. This uses a weighted calculation to assign patients a complication severity score out of 100, based on the number and severity of each recorded postoperative complication.

No power calculation was performed for this retrospective cohort study. Differences in patient characteristics, Comprehensive Complication Index and PhenoAge (and its individual components) between patients with and without postoperative complications were compared using either χ² or Mann–Whitney U tests, as indicated. Quantitative association between PhenoAge and Comprehensive Complication Index was assessed by Pearson's correlation coefficient (r). All analyses were conducted by the MedCalc® software (MedCalc Software Ltd, Ostend, Belgium) and a p value < 0.05 without Bonferroni adjustment was taken as statistically significant.

Of the 49 patients included in our analysis, 28 experienced a postoperative complication within 30 days of surgery (Table 1). Hospital duration of stay was longer in patients who experienced a postoperative complication (median (IQR [range]) 13 (8–18 [1–73]) vs. 6 (5–9 [2–17]) days, p < 0.001). Patients who experienced a postoperative complication were statistically more likely to have a biological age more than their chronological age than those who did not experience a complication (22 vs. 10 patients, respectively; p = 0.024) (Fig. 1a). PhenoAge (Fig. 1b) was significantly correlated with the severity of postoperative complications (r = 0.414, p = 0.003). Baseline albumin was lower in patients who experienced a postoperative complication compared with those who had no complications (median (IQR [range]) 34 g.l^-1 (28–37 [13–42]) vs. 39 g.l^-1 (37–40 [32–41], respectively; p < 0.001)). Baseline red cell distribution width was higher in patients who experienced a postoperative complication compared with those who had no complications (median (IQR [range]) 15% (13–16 [11–28]) vs. 13% (12–13 [11–16]), respectively; p = 0.007).

Table 1. Characteristics of patients with and without complications after major cancer surgery. Values are median (IQR [range]) or number.

Variable	With complications	Without complications	p value
	n = 28	n = 21
Chronological age; y	65 (47–76 [31–86])	54 (49–64 [33–85])	0.225
Sex; male	13	15	0.080
BMI; kg.m-2	25 (22–30 [18–35])	28 (25–32 [17–45])	0.079
Duration of hospital stay; days	13 (8–18 [1–73])	6 (5–9 [2–17])	< 0.001
Days alive and at home at 90 days after surgery	76 (66–82 [0–89])	83 (81–85 [65–88])	< 0.001
Comprehensive Complication Index; (0–100)	22 (13–33 [9–100])	0 (0–0 [0–0])	< 0.001
PhenoAge; y	74 (57–90 [33–116])	53 (48–71 [30–90])	0.002
PhenoAge in excess of chronological age; y	11 (1–26 [-10–47])	-1 (-3 to 6 [-8–13])	0.006
PhenoAgeAccel	22	10	0.024
Components of the PhenoAge
Albumin; g.l-1	34 (28–37 [13–42])	39 (37–40 [32–41])	< 0.001
Creatinine; umol.l-1	69 (61–80 [50–105])	71 (66–82 [64–170])	0.179
Glucose; mmol.l-1	5.8 (5.1–7.3 [3.6–14.6])	5.5 (5.2–6.1 [3.9–7.7])	0.178
C-reactive protein; mg.dl-1	5 (1–26 [0–237])	4 (2–9 [1–105])	0.951
Lymphocyte count; %	23 (15–30 [5–51])	23 (16–29 [9–48])	0.801
Mean cell volume; fl	91 (86–95 [79–111])	92 (86–96 [80–102])	0.887
Red cell distribution width; %	15 (13–16 [11–23])	13 (12–13 [11–16])	0.007
Alkaline phosphatase; U/l	78 (72–95 [34–386])	82 (65–98 [42–148])	0.903
White cell count; 103 cells.ml-1	7.6 (5.7–8.2 [3.2–18.2])	5.9 (4.0–7.6 [2.5–11.7])	0.090

Figure 1

Open in figure viewerPowerPoint

Association between PhenoAge and postoperative complications. (a) Higher incidence of PhenoAgeAccel in patients with postoperative complications (p = 0.006^#); (b) PhenoAge (y) was positively correlated with severity of postoperative complications (r = 0.414, p = 0.003*). Continuous blue line indicates the linear regression line and the dotted lines and blue shaded area indicate the 95%CI of the regression line. *Pearson's correlation. ^#Mann–Whitney U test.

In this proof-of-concept analysis, we found that pre-operative biological ageing is related to postoperative complications after cancer surgery. It is important to consider the limitations of this study, namely the small sample size, retrospective design and potential confounding from patient inclusion, due to the nature of patients with only full baseline PhenoAge data being included. It is possible that this represents a higher risk cohort in whom more comprehensive pre-operative investigations were undertaken. Nonetheless, given the importance of identification and modulation of pre-operative risk to predicting clinical outcomes of surgical patients, the correlations between different PhenoAge, some of its components and severity of postoperative complications may suggest that markers of biological age could be a useful tool for prediction of postoperative complications after surgery. Previous studies showed that biological age is a dynamic epigenetic clock that can be ‘dialled forward’ (by stress, disease or surgery) or ‘backward’ (with interventions such as exercise and nutritional supplementation) [8]. Whether PhenoAge in patients undergoing cancer surgery can be improved with prehabilitation, and, if so, whether the responders will have a reduced risk of developing postoperative complications, remains uncertain and this requires further investigation.