Predictive value of perioperative procalcitonin, C reactive protein and high-sensitivity C reactive protein for the risk of postoperative complications after non-cardiac surgery in elderly patients: a nested case–control study

Abstract

Objective

Little is known about the correlation between perioperative concentrations of inflammatory biomarkers and postoperative complications. This study explored whether the plasma concentrations and perioperative changes of procalcitonin (PCT), C reactive protein (CRP) and high-sensitivity CRP (hsCRP) could predict the risk of postoperative morbidity in elderly patients undergoing elective non-cardiac surgery.

Design

A nested case–control study.

Setting

A tertiary hospital in China.

Participants

A total of 498 patients aged ≥65 years from a prospective cohort who underwent elective non-cardiac surgery between June 2020 and April 2021.

Primary outcome measures

The primary outcomes were the efficacy of plasma concentrations of PCT, CRP and hsCRP in predicting the risk of Clavien-Dindo Classification (CDC) ≥grade 3 and major complications. The major complications included mortality, an intensive care unit stay length >24 hour, cardiovascular events, acute kidney injury, postoperative cognitive dysfunction and infections.

Results

For major complications, the area under the curve (AUC) (95% CI) of PCT-24 hours, PCT change and PCT change rate were 0.750 (0.698 to 0.803), 0.740 (0.686 to 0.795) and 0.711 (0.651 to 0.771), respectively. The AUC (95% CI) of CRP-24 hours, CRP change, CRP change rate and hsCRP baseline were 0.835 (0.789 to 0.881), 0.818 (0.770 to 0.867), 0.691 (0.625 to 0.756) and 0.616 (0.554 to 0.678), respectively. For complications ≥CDC grade 3, the AUC (95% CI) of PCT-24 hours, PCT change and PCT change rate were 0.662 (0.543 to 0.780), 0.643 (0.514 to 0.772) and 0.627 (0.494 to 0.761), respectively. The AUC (95% CI) of CRP-24 hours and hsCRP baseline were 0.649 (0.527 to 0.771) and 0.639 (0.530 to 0.748), respectively.

Conclusions

PCT-24 hours, CRP-24 hours, the change of perioperative PCT and CRP were valuable predictors of major complications occurring within 30 days after non-cardiac surgery in the elderly.

Trial registration number

China Clinical Trial Registry: ChiCTR1900026223.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The study focused on the relationship between perioperative concentrations of inflammatory biomarkers and postoperative complications after non-cardiac surgery in elderly patients.

• The study enrolled a relatively large cohort of elderly patients undergoing elective non-cardiac surgery, and were followed up for 30 days after surgery.

• The study only recruited patients whose surgeries began before 10:00 am, so the reported incidence of postoperative complications might be affected by the surgery time.

• A large cohort investigation in multiple centres is required to evaluate the predictive effect of perioperative concentrations of procalcitonin, C reactive protein and high-sensitivity CRP on postoperative morbidity in elderly patients who underwent elective non-cardiac surgery.

Introduction

The proportion of people aged over 65 years in the world is expected to reach 16% in 2050.¹ Among a total of more than 300 million surgeries in the world every year, patients aged 65 years and over account for more than 30% of cases.² The increasing rate of surgeries in aged patients has exceeded the rapid ageing rate of the population.³ Compared with adults, elderly patients are more prone to postoperative complications due to age-related degenerative physiological characteristics and increasing coexisting morbidities.^{3 4} At least one postoperative complication has been reported in approximately 7% of patients who underwent non-cardiac surgery.⁵ Moreover, with age, postoperative complications are becoming more common.^{6 7}

The risk assessment of postoperative complications is a prerequisite for intervention and prevention. At present, the risk assessment methods are mainly models and scales, and the core indicators are mostly preoperatively coexisting morbidities of patients, which do not truly reflect the consequences of anaesthesia and surgery, and neither approach contributes to dynamic monitoring or assessment during surgery.⁸ Biomarkers can reflect tissue injury and organ dysfunction, which can independently predict postoperative complications or increase the clinical prognostic information, which has become a hot topic of research in recent years. However, some biomarkers exhibit tissue specificity, such as B-type natriuretic peptide and neutrophil gelatinase-associated lipocalin, which have good accuracy in predicting specific organ complications, but there are some limitations in assessing the systemic risk of elderly patients.^{9 10}

Inflammation is one of the main pathophysiological changes that occur during the perioperative period.¹¹ Perioperative inflammatory responses are associated with multiple postoperative complications, including infections,¹² pulmonary complications,¹³ postoperative cognitive dysfunction (POCD)¹⁴ and kidney injury.¹⁵ Due to physiological degeneration and multiple organ diseases, elderly patients usually present with a chronic inflammatory background.¹⁶ After the unavoidable stimulation that is induced by anaesthesia and surgery, the inflammatory response pattern may be different from that of younger patients.

Inflammation is a highly significant risk factor for both morbidity and mortality in elderly people.¹⁷ Both procalcitonin (PCT) and C reactive protein (CRP) are non-specific proteins associated with systemic inflammation, and can aid in the identification of the infection process, characterise the severity of diseases and guide treatment and risk stratification under a variety of clinical conditions.¹⁸ In recent years, it has been found that these factors can be used as predictors of morbidity and mortality in cardiovascular disease,¹⁹ as well as for the prognosis of cardiac surgery and organ transplantation.^{18 20} High-sensitivity C reactive protein (hsCRP) is also widely used for predicting cardiovascular risk and is usually included in risk prediction tools and current cardiovascular disease risk assessment guidelines.²¹ However, there is a paucity of studies on the correlation between these biomarkers and the postoperative complications of non-cardiac surgery in elderly patients.

Therefore, the aim of the present study was to investigate whether the plasma concentration and perioperative changes in plasma PCT, CRP and hsCRP concentrations could predict the risk of postoperative morbidity in elderly patients who underwent elective non-cardiac surgery.

Methods

Ethics approvals and registration

This was a single-centre, nested case–control study from a prospective observational cohort conducted according to Strengthening the Reporting of Observational Studies in Epidemiology guidelines. All procedures performed were in accordance with the Helsinki Declaration. The protocol of the study was approved (No. 199, 2020) by the Ethics Committee on Biomedical Research of West China Hospital of Sichuan University on 25 May 2020, and was registered at the China Clinical Trial Registry (ChiCTR1900026223, https://www.chictr. org.cn).

Patients

From a prospective cohort of aged patients scheduled to receive surgery under general anaesthesia, we consecutively recruited patients whose surgeries began before 10:00 am in West China Hospital of Sichuan University from June 2020 to April 2021. Before surgery, trained investigators identified and contacted suitable patients according to the following criteria: (1) patients aged ≥65 years; (2) no limitation of sex; (3) patients due to undergo elective non-cardiac surgery under general anaesthesia, scheduled to begin before 10:00 am; (4) American Society of Anesthesiologists (ASA) physical status I–IV; (5) patients had understood the protocol to be carried out and provided written signed informed consent. The exclusion criteria were: (1) patients with major preoperative coexisting morbidities that may overlap with complications after surgery including acute kidney injury (AKI), acute coronary syndrome, stroke, pneumonia, delirium and so on; (2) patients with haematological diseases or immunodeficiencies; (3) patients with mental illness or an inability to provide written consent or cooperate during the study; and (4) patients with a history of allergy to anaesthetics. All patients were told that their peripheral blood samples would be collected and analysed during the study. The patients could freely withdraw from the study at any time for any reason.

Anaesthesia procedure

All patients underwent a standard preoperative assessment, including their medical history, physical examination, laboratory blood tests, chest X-ray and a 12-lead ECG. The medical team decided whether further examinations were required. The day before surgery, the responsible anaesthesiologist evaluated the patient to determine the most effective anaesthesia regimen to use.

In the operation room, the patient was routinely monitored with electrocardiography, non-invasive or invasive arterial blood pressure, pulse oximetry and the bispectral index (BIS). Other measurements such as central venous pressure were performed if necessary. After endotracheal intubation, the end tidal CO₂ partial pressure (P_ETCO₂) and airway pressure were continuously monitored. General anaesthesia was intravenously induced with 1.5–2.5 mg/kg propofol, 0.3 µg/kg sufentanil, and 0.1–0.2 mg/kg cisatracurium or 0.08–0.1 mg/kg vecuronium bromide. After tracheal intubation, a ventilation setting protocol of 6–8 mL/kg tidal volume and 10–16 beats/min respiratory rates were established to maintain a P_ETCO₂ of 35–45 mmHg. The opioids (remifentanil and sufentanil), muscle relaxants (cisatracurium or vecuronium bromide), sevoflurane, desflurane and propofol were used to maintain anaesthesia in the BIS range of 40–60. The infusion of intraoperative fluids and blood products, and the management of vasopressors were determined according to the intraoperative arterial blood gas results and the clinical judgement of the anaesthesiologist. After surgery, the patient was extubated then transferred to a postanaesthesia care unit, or directly to the intensive care unit (ICU), according to the decision of the anaesthesiologist.

PCT, CRP and hsCRP testing

All patients underwent venous blood sampling twice, to measure plasma concentrations of PCT, CRP and hsCRP before surgery, in the operation room for PCT, CRP, hsCRP baselines, and 24 hours after surgery for PCT-24 hours, CRP-24 hours and hsCRP-24 hours determinations. The blood samples were collected in EDTA evacuated tubes and centrifuged immediately, and then measured using a Finecare III Plus (FS-205) which is an immunofluorescence automatic analyser (GuangzhouWanfu Biotechnology Co, Guangzhou, China). The detection lower limits of PCT, CRP and hsCRP concentrations were 0.05 ng/mL, 2.5 mg/L and 0.25 mg/L, respectively. If the detection concentration of the blood sample was less than its detection lower limit, it was recorded as the lowest value that could be detected. The upper limit of hsCRP detection by the analyser was 5 mg/L. If the detection concentrations in the blood sample was >5 mg/L, it was recorded as 5 mg/L. The normal reference ranges were 0–0.5 ng/mL, 0–10 mg/L, 0–1 mg/L for PCT, CRP and hsCRP, respectively.

Data collection

The basic demographic, clinical and laboratory data of patients were extracted from the hospital’s electronic medical records system. The demographics included age, sex, weight, height, body mass index and smoking status. The comorbidities mainly included hypertension, ischaemic heart disease, arrhythmias, valvular heart disease, diabetes mellitus, chronic obstructive pulmonary disease, cerebrovascular disease, liver cirrhosis, renal dysfunction and so on. The ASA physical status, New York Heart Association (NYHA) cardiac functional class, motion equivalent, type of surgery, and the surgical and anaesthesia times were also carefully recorded.

Once patients signed consent forms, they were anonymised with an identification code. The preoperative assessment and postoperative follow-up were performed by different trained investigators, who were blind to each other’s data.

The severities of any complications were categorised using a modified Clavien-Dindo Classification (CDC) scheme.²² The incidence of moderate to severe complications were defined as CDC ≥grade 3. The major complications included mortality, an ICU stay length >24 hours, cardiovascular events, AKI, POCD and infections. Mortality was defined as disease-related or surgery-related deaths within 30 days after surgery. AKI was defined according to kidney disease improving global outcomes (KDIGO) criteria (2012).²³ Cardiovascular events included the new onset or aggravation of congestive heart failure, acute myocardial infarction or arrhythmias.²⁴ POCD was diagnosed by International Study Group of Postoperative Cognitive Dysfunction (ISPOCD) method.²⁵ The postoperative complications were monitored by trained researchers by interviews during hospitalisation or telephone calls after hospital discharge, to determine unequivocally whether there were any postoperative complications. The primary outcome of the study was the efficacy of plasma concentration of PCT, CRP, hsCRP for predicting the risk of CDC ≥grade 3 and major complications.

Statistical analysis

The sample size calculation was based on comparing the area under the curve (AUC) of the receiver operating characteristic (ROC) curves using MedCalc (V.19). Based on our previous study, assuming an incidence of complications ≥CDC grade 3 of 4% and a moderately good AUC of 0.7 for blood PCT, CRP or hsCRP concentrations, a sample size of 472 patients would yield a 90% power to detect clinically relevant differences in AUC values (two-sided alpha of 0.05). To account for 10% of patients who may be lost to follow-up, we aimed to enrol a total of 491 patients in the study.

Statistical analysis was performed using SPSS V.15 (SPSS, Chicago, Illinois, USA). All data from different groups were verified for normality and homogeneity of variance using Kolmogorov-Smirnov and Brown-Forsythe tests before analysis. The Wilcoxon test was used for comparisons between two groups. ROC curve analysis and AUC results were used to explore the predictive value of perioperative PCT, CRP and hsCRP concentrations for postoperative complications. An AUC between 0.5 and 0.6 indicated bad accuracy of the diagnostic test. An AUC between 0.6 and 0.7 indicated sufficient accuracy, between 0.7 and 0.8 good accuracy, between 0.8 and 0.9 very good accuracy and an AUC >0.9 indicated excellent accuracy of the diagnostic test.²⁶ The cut-off values were determined based on the Youden index. All statistical tests were two-sided, and significance was assumed at a p value <0.05.

Patient and public involvement

Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Results

In total, 498 patients were included in the final analysis between June 2020 and April 2021, as shown in online supplemental figure 1. The demographic characteristics and preoperative comorbidities of the patients, and information on their surgery and anaesthesia are shown in table 1.

Table 1.

The clinical characteristics of patients (n=498)

Variable	Value
Characteristics
Age (years)	71.8±5.2
Sex
Male	328 (65.86%)
Female	170 (34.14%)
Height (cm)	161.2±8.1
Weight (kg)	60.1±9.9
BMI	23.1±3.3
Current or recent smoker	72 (14.46%)
Comorbidities
Hypertension	200 (40.16%)
I	68 (13.65%)
II	80 (16.06%)
III	52 (10.44%)
Ischaemic heart disease	33 (6.63%)
Arrhythmias	20 (4.02%)
Valvular heart disease	8 (1.61%)
Diabetes mellitus	206 (41.37%)
COPD	17 (3.41%)
Cerebrovascular disease	12 (2.41%)
ASA physical status
II	225 (45.18%)
III	272 (54.62%)
IV	1 (0.20%)
NYHA cardiac functional class
1	292 (58.63%)
2	191 (38.35%)
3	15 (3.01%)
Motion equivalent
<3 MET	68 (13.65%)
3–6 MET	354 (1.08%)
>6 MET	76 (15.26%)
Type of surgery
Orthopaedic	79 (15.86%)
Hepatobiliary	119 (23.90%)
Gastrointestinal	154 (30.92%)
Pancreatic	45 (9.03%)
Urological	71 (14.26%)
Thoracic	30 (6.02%)
Open surgery	305 (61.24%)
Endoscopy surgery	193 (38.76%)
Surgery time (min)	182±90
Anaesthesia time (min)	253±102

Values are expressed as the mean±SD or as numbers (percent).

ASA, American Society of Anesthesiologists; BMI, body mass index; COPD, chronic obstructive pulmonary disease; MET, metabolic equivalent; NYHA, New York Heart Association.

The concentrations of PCT (0.05 (0.05, 0.05) ng/mL vs 0.41 (0.20, 1.24) ng/mL, p<0.001), CRP (2.50 (2.50, 2.50) mg/L vs 35.55 (21.70, 57.98) mg/L, p<0.001) and hsCRP (1.20 (0.25, 3.80) mg/L vs 5.00 (5.00, 5.00) mg/L, p<0.001) 24 hours after operation were significantly higher than those before operation.

The postoperative complications of 26 (5.22%) patients were ≥CDC grade 3. There were 103 (20.68%) patients with major complications, 2 patients (0.40%) died, 13 (2.61%) stayed in ICU for >24 hours, 18 (3.61%) had cardiovascular events, 10 (2.00%) had AKI, 15 (3.01%) POCD and 103 (20.68%) with infections.

As listed in table 2, for patients who developed complications ≥CDC grade 3, the baseline concentrations of CRP (p=0.032) and hsCRP (p=0.015) were higher than those of negative patients. At 24 hours after surgery, the concentrations of PCT (p=0.005) and CRP (p=0.011) in complications ≥CDC grade 3 patients were significantly higher than those of negative patients, while there was no difference in the concentrations of hsCRP between the two groups (p=0.556). Correspondingly, the PCT change (PCT-24 hours − PCT baseline) and PCT change rate ((PCT-24 hours − PCT baseline)/PCT baseline × 100%) were also higher in positive patients than those in negative patients (PCT change: p=0.014; PCT change rate: p=0.029). However, no statistical differences were found in the changes of concentrations of CRP (CRP change: p=0.086; CRP change rate: p=0.764).

Table 2.

The comparisons of perioperative concentrations of PCT, CRP, hsCRP between positive and negative patients with complications ≥CDC grade 3 or major complications

Variable	Positive patients	Negative patients	P value
Complications ≥CDC grade 3	(n=26)	(n=472)
PCT baseline (ng/mL)	0.05 (0.05, 0.05)	0.05 (0.05, 0.05)	0.221
PCT-24 hours (ng/mL)	1.61 (0.26, 2.60)	0.40 (0.19, 1.15)	0.005
PCT change (ng/mL)	1.34 (0.21, 2.55)	0.33 (0.14, 1.05)	0.014
PCT change rate (%)	22.80 (2.65, 49.35)	6.40 (2.20, 17.30)	0.029
CRP baseline (mg/L)	2.50 (2.50, 11.58)	2.50 (2.50, 2.50)	0.032
CRP-24 hours (mg/L)	59.65 (29.00, 92.65)	34.90 (21.35, 55.98)	0.011
CRP change (mg/L)	41.25 (20.25, 67.88)	29.65 (17.23, 52.00)	0.086
CRP change rate (%)	9.80 (3.40, 16.81)	10.12 (4.81, 17.24)	0.764
hsCRP baseline (mg/L)	2.70 (0.85, 5.00)	1.20 (0.25, 3.60)	0.015
hsCRP-24 hours (mg/L)	5.00 (5.00, 5.00)	5.00 (5.00, 5.00)	0.556
hsCRP change (mg/L)	2.25 (0.00, 3.85)	3.80 (1.20, 4.75)	0.011
hsCRP change rate (%)	0.82 (0.00, 3.67)	3.17 (0.35, 19.00)	0.011
Major complications	(n=103)	(n=395)
PCT baseline (ng/mL)	0.05 (0.05, 0.05)	0.05 (0.05, 0.05)	0.022
PCT-24 hours (ng/mL)	1.43 (0.54, 2.85)	0.35 (0.17, 0.82)	<0.001
PCT change (ng/mL)	1.25 (0.49, 2.62)	0.28 (0.11, 0.76)	<0.001
PCT change rate (%)	19.60 (7.40, 48.80)	5.20 (2.15, 12.64)	<0.001
CRP baseline (mg/L)	2.50 (2.50, 2.50)	2.50 (2.50, 2.50)	0.002
CRP-24 hours (mg/L)	71.70 (50.50, 97.00)	30.30 (19.10, 45.80)	<0.001
CRP change (mg/L)	63.00 (40.70, 86.90)	26.10 (15.20, 41.60)	<0.001
CRP change rate (%)	18.72 (8.00, 30.88)	9.00 (4.47, 14.60)	<0.001
hsCRP baseline (mg/L)	2.50 (0.60, 5.00)	1.00 (0.25, 3.10)	<0.001
hsCRP-24 hours (mg/L)	5.00 (5.00, 5.00)	5.00 (5.00, 5.00)	0.345
hsCRP change (mg/L)	2.50 (0, 4.40)	3.80 (1.70, 4.75)	<0.001
hsCRP change rate (%)	1.00 (0, 7.33)	3.55 (0.56, 19.00)	<0.001

Values are expressed as median (IQR).

CDC, Clavien-Dindo Classification; CRP, C reactive protein; hsCRP, high-sensitivity C reactive protein; PCT, procalcitonin.

For major complications, the concentrations of PCT and CRP in positive patients were higher than those in negative patients at baseline and 24 hours after surgery (PCT baseline: p=0.022; PCT-24 hours: p<0.001; CRP baseline: p=0.002; CRP-24 hours: p<0.001, table 2). Correspondingly, the changes of concentrations of PCT and CRP were also higher in positive patients than in negative patients. The concentrations of hsCRP baseline were also higher in positive patients than in negative patients (p<0.001, table 2). Interestingly, the changes of concentrations of hsCRP were lower in positive patients than those in negative patients (table 2). This may be because the upper limit of the concentrations of hsCRP detection by the analyser was 5 mg/L, which limited the response to higher values. In addition, the hsCRP baseline of positive patients was inherently high, so the rise of positive patients could not be accurately reflected.

For patients who stayed in ICU for more than 24 hours after surgery, the concentration of basic CRP was higher than that for corresponding negative patients (p=0.010, table 3). The concentration of CRP-24 hours and perioperative changes of CRP in patients with AKI were significantly higher than those in negative patients (CRP-24 hours: p=0.006; CRP change: p=0.009, table 3). Because the patients with postoperative infections were similar to those with major complications, perioperative concentrations and changes of PCT, CRP and hsCRP were consistent with those patients with major complications (table 3). Further, the various postoperative infections in patients were differentiated, and classified as a wound infection (n=21), pneumonia (n=84), urinary tract infection (n=3) and antibiotic empiric therapy (n=83). The concentrations and changes of PCT, CRP and hsCRP are shown in online supplemental table 1. The concentrations and changes of PCT, CRP and hsCRP in patients with cardiovascular events and POCD were not statistically different from those in negative patients. These results are given in online supplemental table 2.

Table 3.

The comparisons of perioperative concentrations of PCT, CRP, hsCRP between positive and negative patients with an ICU stay length >24 hours or AKI

Variable	Positive patients	Negative patients	P value
ICU stay length >24 hours	(n=13)	(n=485)
PCT baseline (ng/mL)	0.05 (0.05, 0.75)	0.05 (0.05, 0.05)	0.204
PCT-24 hours (ng/mL)	1.52 (0.18, 2.16)	0.41 (0.20, 1.21)	0.364
PCT change (ng/mL)	0.91 (0.05, 2.11)	0.34 (0.14, 1.12)	0.510
PCT change rate (%)	7.60 (0.58, 40.70)	6.50 (2.36, 18.09)	0.806
CRP baseline (mg/L)	2.50 (2.50, 21.65)	2.50 (2.50, 2.50)	0.010
CRP-24 hours (mg/L)	61.80 (14.60, 98.10)	35.30 (21.75, 56.35)	0.156
CRP change (mg/L)	53.40 (12.10, 68.95)	29.70 (17.60, 52.15)	0.382
CRP change rate (%)	5.76 (0.59, 17.52)	10.20 (4.98, 17.11)	0.099
hsCRP baseline (mg/L)	2.30 (0.58, 5.00)	1.20 (0.25, 3.70)	0.102
hsCRP-24 hours (mg/L)	5.00 (5.00, 5.00)	5.00 (5.00, 5.00)	0.171
hsCRP change (mg/L)	0.40 (0, 4.18)	3.70 (1.15, 4.75)	0.058
hsCRP change rate (%)	0.44 (0 10.79)	3.17 (0.32, 19.00)	0.062
AKI	(n=10)	(n=488)
PCT baseline (ng/mL)	0.05 (0.05, 0.63)	0.05 (0.05, 0.05)	0.529
PCT-24 hours (ng/mL)	1.46 (0.26, 4.69)	0.41 (0.19, 1.22)	0.070
PCT change (ng/mL)	1.41 (0.21, 4.56)	0.34 (0.13, 1.13)	0.057
PCT change rate (%)	22.25 (44.25, 46.70)	6.40 (2.20, 18.14)	0.051
CRP baseline (mg/L)	2.50 (2.50, 17.05)	2.50 (2.50, 2.50)	0.333
CRP-24 hours (mg/L)	75.30 (36.38, 105.98)	34.90 (21.55, 56.38)	0.006
CRP change (mg/L)	59.65 (33.88, 100.08)	29.70 (17.23, 52.18)	0.009
CRP change rate (%)	13.38 (4.48, 40.03)	9.94 (4.79, 17.06)	0.300
hsCRP baseline (mg/L)	1.90 (0.51, 5.00)	1.20 (0.25, 3.70)	0.338
hsCRP-24 hours (mg/L)	5.00 (5.00, 5.00)	5.00 (5.00, 5.00)	0.632
hsCRP change (mg/L)	3.10 (0, 4.49)	3.70 (0.93, 4.75)	0.393
hsCRP change rate (%)	1.70 (0, 10.25)	3.17 (0.32, 19.00)	0.372
Infections	(n=103)	(n=395)
PCT baseline (ng/mL)	0.05 (0.05, 0.05)	0.05 (0.05, 0.05)	0.022
PCT-24 hours (ng/mL)	1.43 (0.54, 2.85)	0.35 (0.17, 0.82)	<0.001
PCT change (ng/mL)	1.25 (0.49, 2.62)	0.28 (0.11, 0.76)	<0.001
PCT change rate (%)	19.60 (7.40, 48.80)	5.20 (2.15, 12.64)	<0.001
CRP baseline (mg/L)	2.50 (2.50, 2.50)	2.50 (2.50, 2.50)	0.002
CRP-24 hours (mg/L)	71.70 (50.50, 97.00)	30.30 (19.10, 45.80)	<0.001
CRP change (mg/L)	63.00 (40.70, 86.90)	26.10 (15.20, 41.60)	<0.001
CRP change rate (%)	18.72 (8.00, 30.88)	9.00 (4.47, 14.60)	<0.001
hsCRP baseline (mg/L)	2.50 (0.60, 5.00)	1.00 (0.25, 3.10)	<0.001
hsCRP-24 hours (mg/L)	5.00 (5.00, 5.00)	5.00 (5.00, 5.00)	0.345
hsCRP change (mg/L)	2.50 (0, 4.40)	3.80 (1.70, 4.75)	<0.001
hsCRP change rate (%)	1.00 (0, 7.33)	3.55 (0.56, 19.00)	<0.001

Values are expressed as median (IQR).

AKI, acute kidney injury; CRP, C reactive protein; hsCRP, high-sensitivity C reactive protein; ICU, intensive care unit; PCT, procalcitonin.

The ROC analysis gave a prediction of perioperative concentrations of PCT, CRP and hsCRP for 30-day complications. For complications ≥CDC grade 3, the AUC (95% CI) of PCT-24 hours, PCT change and PCT change rate were 0.662 (0.543, 0.780) (p=0.006), 0.643 (0.514, 0.772) (p=0.014) and 0.627 (0.494, 0.761) (p=0.029), respectively. The cut-off concentrations of the PCT-24 hours, PCT change and PCT change rate were 1.115 ng/mL, 0.620 ng/mL and 21.5%, respectively, with the best combined sensitivities and specificities. The AUC (95% CI) of the CRP-24 hours and hsCRP baseline were 0.649 (0.527, 0.771) (p=0.011) and 0.639 (0.530, 0.748) (p=0.017), respectively. The corresponding cut-off concentrations were 57.300 mg/L and 2.250 mg/L. For major complications and infections, the AUC (95% CI) of PCT-24 hours, PCT change and PCT change rate were 0.750 (0.698, 0.803) (p<0.001), 0.740 (0.686, 0.795) (p<0.001) and 0.711 (0.651, 0.771) (p<0.001), respectively. The AUC (95% CI) of CRP baseline, CRP-24 hours, CRP change, CRP change rate and hsCRP baseline were 0.570 (0.505, 0.635) (p=0.028), 0.835 (0.789, 0.881) (p<0.001), 0.818 (0.770, 0.867) (p<0.001), 0.691 (0.625, 0.756) (p<0.001) and 0.616 (0.554, 0.678) (p<0.001), respectively. For AKI, the AUC (95% CI) of CRP-24 hours and CRP change were 0.754 (0.591, 0.918) (p=0.006) and 0.741 (0.585, 0.897) (p=0.009), respectively. The results of ROC analyses are shown in figure 1 and table 4. The ROC analysis of the predictive value of perioperative concentrations of PCT, CRP and hsCRP for 30-day various infections are shown in online supplemental table 3, figure 2.

Figure 1.

Results of ROC analysis for perioperative PCT, CRP and hsCRP for predict of complications within 30 days after surgery. The perioperative PCT, CRP and hsCRP concentrations for predicting complications ≥CDC grade 3 (A); major complications (B); an ICU stay length >24 hours (C); AKI (D); and infections (E). AKI, acute kidney injury; CDC, Clavien-Dindo Classification; CRP, C reactive protein; hsCRP, high-sensitivity CRP; ICU, intensive care unit; PCT, procalcitonin; ROC, receiver operating characteristic.

Table 4.

The predict of perioperative concentrations of PCT, CRP and hsCRP for 30-day complications

Variable	AUC (95% CI)	Cut-off value	Sensitivity	Specificity	P value
Complications ≥CDC grade 3
PCT-24 hours	0.662 (0.543, 0.780)	1.115	0.615	0.748	0.006
PCT change	0.643 (0.514, 0.772)	0.620	0.692	0.657	0.014
PCT change rate	0.627 (0.494, 0.761)	21.530	0.538	0.788	0.029
CRP baseline	0.588 (0.463, 0.712)	8.900	0.308	0.114	0.132
CRP-24 hours	0.649 (0.527, 0.771)	57.300	0.538	0.763	0.011
hsCRP baseline	0.639 (0.530, 0.748)	2.250	0.654	0.657	0.017
Major complications
PCT baseline (ng/mL)	0.541 (0.477, 0.606)	0.115	0.165	0.922	0.195
PCT-24 hours (ng/mL)	0.750 (0.698, 0.803)	0.436	0.806	0.630	<0.001
PCT change (ng/mL)	0.740 (0.686, 0.795)	0.455	0.786	0.653	<0.001
PCT change rate (%)	0.711 (0.651, 0.771)	9.100	0.728	0.671	<0.001
CRP baseline (mg/L)	0.570 (0.505, 0.635)	5.100	0.311	0.828	0.028
CRP-24 hours (mg/L)	0.835 (0.789, 0.881)	47.900	0.806	0.767	<0.001
CRP change (mg/L)	0.818 (0.770, 0.867)	46.500	0.718	0.808	<0.001
CRP change rate (%)	0.691 (0.625, 0.756)	18.700	0.505	0.858	<0.001
hsCRP baseline (mg/L)	0.616 (0.554, 0.678)	1.550	0.621	0.600	< 0.001
ICU stay length >24 hours
CRP baseline	0.648 (0.474, 0.821)	5.950	0.462	0.833	0.069
AKI
CRP-24 hours	0.754 (0.591, 0.918)	64.650	0.700	0.812	0.006
CRP change	0.741 (0.585, 0.897)	32.500	0.900	0.542	0.009
Infections
PCT baseline (ng/mL)	0.541 (0.477, 0.606)	0.115	0.165	0.922	0.195
PCT-24 hours (ng/mL)	0.750 (0.698, 0.803)	0.436	0.806	0.630	<0.001
PCT change (ng/mL)	0.740 (0.686, 0.795)	0.455	0.786	0.653	<0.001
PCT change rate (%)	0.711 (0.651, 0.771)	9.100	0.728	0.671	<0.001
CRP baseline (mg/L)	0.570 (0.505, 0.635)	5.100	0.311	0.828	0.028
CRP-24 hours (mg/L)	0.835 (0.789, 0.881)	47.900	0.806	0.767	<0.001
CRP change (mg/L)	0.818 (0.770, 0.867)	46.500	0.718	0.808	<0.001
CRP change rate (%)	0.691 (0.625, 0.756)	18.700	0.505	0.858	<0.001
hsCRP baseline (mg/L)	0.616 (0.554, 0.678)	1.550	0.621	0.600	<0.001

AKI, acute kidney injury; AUC, area under the curve; CDC, Clavien-Dindo Classification; CRP, C reactive protein; hsCRP, high-sensitivity C reactive protein; ICU, intensive care unit; PCT, procalcitonin.

Discussion

Few studies have been carried out on inflammatory indicators and postoperative prognosis of elderly patients. In the present study, the incidence of postoperative major complications occurring within 30 days was 20.68% in elderly patients who underwent non-cardiac surgery. PCT-24 hours, CRP-24 hours, the change of perioperative PCT and CRP were valuable predictors of major complications and infections. For postoperative AKI, CRP-24 hours and the change of perioperative CRP were valuable predictors.

In the clinic, the PCT concentration is mainly used as a biomarker of bacterial infection, and as an auxiliary tool to diagnose lower respiratory tract and intraperitoneal infection, bacteraemia and sepsis. However, PCT concentrations may also be increased in cirrhosis, pancreatitis, cardiogenic shock, trauma and ischaemic intestinal obstruction.²⁰ Previous studies reported that PCT was also associated with mortality after cardiac surgery^{27 28} and organ transplantation.^{29 30} A high concentration of PCT was also found to be related to various postoperative complications. A high plasma PCT concentration after surgery was associated with perioperative myocardial infarction after cardiac surgery.³¹ Postoperative PCT elevation was a valuable predictor of systemic inflammatory response syndrome, cardiac insufficiency and respiratory insufficiency in patients after cardiac surgery.^{32 33} At present, there is a lack of PCT research on the postoperative complications of non-cardiac surgery in elderly patients. The present study found that PCT 24 hours after surgery and changes in PCT concentrations in the perioperative period were valuable predictors of major complications in elderly patients after non-cardiac surgery.

The CRP concentration was increased in almost all types of surgery,³⁴ and a high CRP concentration was associated with the severity of post-traumatic stress.³⁵ The increase of serum CRP concentration after an operation could identify the risk of postoperative complications in patients with oesophageal cancer, and was also related to the prolongation of ICU hospitalisation and the increase in 1-year mortality.¹⁹ A study involving 1427 patients who underwent abdominal surgery found that patients with major complications had higher CRP concentration after surgery, and that the postoperative CRP concentration was a good predictor of major complications.³⁶ These findings are consistent with the results of our study, namely that CRP-24 hours is an important predictor for major complications. However, the preoperative CRP concentration did not predict postoperative complications, which was opposite to the findings of previous studies. Stiff et al reported that a high preoperative CRP value was a poor prognostic indicator in relation to patient survival after liver cancer surgery.³⁷ This may be because CRP is secreted by the liver in response to stimulation by a variety of inflammatory cytokines.³⁸ For patients undergoing liver cancer surgery, the disease has an impact on preoperative CRP. Balta et al found that elderly patients (83.09±8.52 years) who died within the first 30 days after hip fracture surgery had a higher preoperative CRP concentration compared with survivors, and that the pre-operative CRP concentration was a significant predictor of mortality within 30 days.³⁹ The ages of patients in this study were significantly older than those in our study, and more systemic diseases were combined.

PCT and CRP concentrations are commonly used indicators of infection and are often employed to assist in the diagnosis of infection. Previous studies have reported that PCT and CRP concentrations after cardiac surgery can predict the occurrence of systemic inflammatory response and multiple organ dysfunction syndromes in patients.^40–42 After oesophageal surgery, increased PCT and CRP concentrations are also predictive indicators of major infective complications.⁴³ In the study, PCT-24 hours, CRP-24 hours, changes of PCT and CRP concentrations in the perioperative period were good predictors of infection after non-cardiac surgery in elderly patients. However, preoperative CRP and PCT concentrations did not predict the occurrence of postoperative infection,⁴¹ conclusions consistent with our findings.

Inflammation is an important component of AKI, playing a considerable role in its pathophysiology.⁴⁴ AKI is a common complication after cardiac surgery, with an occurrence rate of about 30%.⁴⁵ The morbidity of postoperative AKI in patients who underwent non-cardiac surgery was significantly lower than for patients who underwent cardiac surgery, even in the elderly. In the study, the morbidity of postoperative AKI was 2.0% (10/498). The incidence of AKI in the ICU has been reported to occur in >50% of patients.⁴⁶ In this study, the incidence of an ICU stay length >24 hours after surgery was also low, being only 2.61% (13/498). This may also be a reason for the low incidence of postoperative AKI in the study. Previous studies have discussed that basic PCT⁴⁷ or PCT-48-hour concentrations after surgery^{48 49} were valuable predictors of AKI in patients who had underwent cardiovascular surgery. Chen et al reported that peak PCT and CRP concentrations were higher in patients with AKI after surgery for acute type A aortic dissection, and both permitted the prediction of AKI after surgery, but it is noteworthy that PCT was a better predictor.⁵⁰ While in our study, we found no statistical differences in PCT baseline, PCT-24 hours and PCT changes in concentrations in elderly non-cardiac surgery patients with and without postoperative AKI. However, the changes in CRP-24 hours and CRP concentrations were good predictors of AKI in elderly patients after non-cardiac surgery.

Inflammation is a characteristic of all stages of atherothrombosis,⁵¹ so various inflammatory factors may be useful clinical markers to predict cardiovascular events. Among them, hsCRP is considered to be the most powerful single factor for predicting cardiovascular events.²¹ Moreover, hsCRP can predict the risk of cardiovascular events in various types of people, including healthy individuals without cardiovascular disease, patients with acute coronary syndrome, patients with stable angina pectoris or patients in the stable stage after myocardial infarction.⁵² In addition, hsCRP is also associated with type 2 diabetes,⁵³ renal disease⁵⁴ and metabolic syndrome.⁵⁵ However, in this study, there was no statistical difference in hsCRP concentrations between patients with and without cardiovascular events or AKI after surgery. We speculate that the possible reason is that the upper limit of the hsCRP concentration detected by the analyser was 5 mg/L, which limits the response to higher values and hides any real changes. In addition, the hsCRP baseline of positive patients was inherently high, so the rise of positive patients could not be accurately reflected.

There are a number of limitations to the present study. First, there is still controversy about the influence of surgery time on the development of postoperative complications.^{56 57} We only enrolled patients whose surgeries began before 10:00 am, so the reported incidence of postoperative complications might be affected by the time that surgery was initiated. Second, this was a single-centre study, and the results need to be verified by a larger cohort investigation in multiple centres. Finally, most of the AUC values were between 0.6 and 0.7, indicating the prediction value was limited. Further multi-centre large sample research will likely increase its prediction efficiency.

Conclusion

After non-cardiac surgery in the elderly, PCT-24 hours, CRP-24 hours, the change of perioperative PCT and CRP concentrations were valuable predictors of major complications and infections within 30 days of surgery. CRP-24 hours and the change of perioperative CRP concentrations were found to be valuable predictors for postoperative AKI.

Data availability statement

Data are available upon reasonable request.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

The protocol of the present study was approved (No. 199, 2020) by the Ethics Committee on Biomedical Research of West China Hospital of Sichuan University.