Lactate Dehydrogenase and Risk of Readmission with Gastric Cancer: A Propensity Score Matching Analysis

Abstract

Purpose

Readmission is one of the measures of quality of care and potential costs. This study aimed to determine whether lactate dehydrogenase (LDH) is associated with an increased risk of 30-day readmission in gastric cancer.

Methods

We performed a retrospective study of patients who underwent radical gastrectomy for gastric cancer at our institution between July 2014 and May 2018. Balanced cohorts were created by propensity score matching (PSM) with a 1:1 ratio to generate the elevated LDH (ELDH) group (n = 151) and the low LDH group (Control) (n = 302). To determine the incidence, causes, and risk factors of 30-day readmission, subgroup analyzes were performed and used to develop an efficient prediction model.

Results

A total of 788 patients met the criteria to be included in the study. The cutoff value for serum LDH was 215.5. After PSM, a total of 302 patients were matched in pairs (ELDH group, n = 151, Control group, n = 151). ELDH levels had a higher risk of readmission (p = 0.005, Odds ratio 3.768, 95% confidence interval 1.493–9.510). The pre-match 30-day readmission rate was 7.2 percent, and common causes of post-match readmission included infection-related symptoms, gastrointestinal symptoms, and gastrointestinal bleeding.

Conclusions

Patients with preoperative ELDH levels, postoperative complications, and high preoperative American Society of Anesthesiologists Scores had a higher risk of readmission 30 days after surgery.

Keywords:

• Serum lactate dehydrogenase
• postoperative complications
• gastric cancer
• readmission
• propensity score matching

Introduction

Gastric cancer has become the third most common cause of social tumor burden in China, with about 480,000 new cases per year [1]. For gastric cancer, surgical resection remains one of the main treatments. With the development of rapid recovery, the time from admission to discharge for surgical patients has been decreasing. This situation has led to concerns that patients leaving the hospital in a vulnerable state will cause unplanned readmissions. Readmissions can significantly increase the medical burden for families and patients, and may impact patient’s postoperative recovery and quality of life [2]. For governments, readmissions add to the financial burden on the health care system. Therefore, for surgical patients, we need appropriate patient perioperative care during hospitalization, as well as assessment, education, and associated follow-up of the patient upon discharge. The United States required a reduction in readmissions as part of the Affordable Care Act of 2010, although the act did not take effect [3, 4]. However, the readmission rate of gastric cancer patients is still 6.6% to 8.7%, which is elevated compared to other tumors [5].

Lactate dehydrogenase (LDH) is closely related to the glycolysis and mitochondrial defects required for tumor growth [6]. The Warburg effect exists in tumor cells indicating that cancer cells preferentially use the anaerobic pathway of glycolysis to generate energy despite the presence of oxygen [7]. LDH is one of the crucial enzymes for anaerobic glycolysis and gluconeogenesis, which catalyzes the redox reaction between propionic acid and L-lactic acid. In addition, LDH is typically cleared by receptor-mediated macrophage endocytosis [8]. When a tumor grows, the immune system within its microenvironment changes accordingly, resulting in insufficient LDH clearance from the tumor [9]. Therefore, patients with high lactate dehydrogenase levels have relatively low overall survival and disease-free survival, and correlation analysis suggests that elevated serum LDH has a negative effect on the prognosis of gastric cancer [10–12], and some scholars recommend LDH as an objective indicator of tumor prognosis [13, 14]. At the same time, some studies have shown that patients with gastric cancer have a higher readmission rate after surgery, and the prognosis associated with it is also worse [15]. Therefore, we believe that LDH is also related to hospital readmissions. However, to the best of our knowledge, although there have been numerous studies on LDH and gastric cancer prognosis, no study has analyzed the correlation between LDH and the 30-day readmission of gastric cancer patients. Therefore, we included LDH in our retrospective study to investigate the association of readmission with LDH in gastric cancer patients.

Materials and methods

Patient

The First Affiliated Hospital of Wenzhou Medical University is a comprehensive hospital integrating medical treatment, teaching, scientific research, and preventive health care. It is responsible for the medical care and treatment of critical and intractable diseases for nearly 30 million people in southern Zhejiang, southern Fujian, and northeastern China. We retrospectively included 927 patients who presented to the Department of Gastrointestinal Surgery of the First Affiliated Hospital of Wenzhou Medical University between July 2014 and May 2018. Our inclusion criteria were: 1) age ≥ 18 years; 2) pathological diagnosis of gastric cancer; 3) surgical treatment at our research center; and 4) readmission at our research center within 30 days of discharge. The exclusion criteria were: 1) patients who died during the initial hospital stay after surgery or within 30 days of discharge; 2) palliative surgery; and 3) history of treatment for cancer. As shown in Figure 1, after examining inclusion and exclusion criteria, we identified 788 eligible patients. After propensity matching, 302 patients were eventually enrolled in the study. All procedures in this study were standard open and laparoscopic gastrectomy performed by experienced surgeons according to guidelines [16, 17]

Figure 1. Block flow chart of patient inclusion. ELDH, Elevated Lactate Dehydrogenase.

Data collection

We collected patient data on three aspects: (1) Basic clinical characteristics: age, gender, Body Mass Index (BMI), American Society of Anesthesiologists (ASA) grade, preoperative comorbidities (using the Charlson Comorbidity Index (CCI) score), preoperative hemoglobin concentration, preoperative plasma albumin concentration, preoperative nutrition risk screening (NRS) score, preoperative serum LDH level, and history of abdominal surgery and sarcopenia; (2) Surgical characteristics: including TNM stage (Tumor Node Metastasis), operation method (laparoscopic or open), whether total gastrectomy was performed, whether combined resection was performed during the operation, operation time (≥210 min), and surgical anastomosis method (Billroth I, Billroth II, Roux-en -Y); and (3) Postoperative outcomes: postoperative complications, postoperative hospital stay, hospitalization costs, readmission within one month after discharge, and readmission causes.

Evaluated parameters

To control for deviations in the study, all subjective assessments were performed by two specially trained independent researchers. They were blinded to basic patient information and other data. Preoperative hypoalbuminemia was defined as preoperative plasma albumin <35 g/L. Preoperative anemia was determined according to the hemoglobin concentration in peripheral blood (hemoglobin concentration <120 g/L in men and <110 g/L in women). The NRS was assessed by NRS 2002 [18], which was completed within 24 hours after the patient was admitted to the hospital, and a score of ≥ 3 was judged as having nutritional risk. The diagnosis of sarcopenia was based on previous studies and defined as: (1) low muscle mass (L3 skeletal muscle index <40.8 cm²/m² in males and <34.9 cm²/m² in females); (2) low muscle strength (handgrip strength <26 kg for males or <18 kg for females); and (3) low muscle performance (usual gait speed <0.8 m/s over 6 m). According to the Clavien-Dindo classification [19], major complications were defined as those that met the criteria of grade II or higher. Readmission was defined as readmission within 30 days of discharge. We chose the 30-day period because the 30-day readmission rate has long been the standard rate reported in previous studies. When a patient was readmitted for more than one reason, the most prominent problem was considered as the reason for readmission.

Propensity score analysis

To reduce the bias due to nonrandom assignment, we used propensity score matching (PSM) to balance the elevated LDH (ELDH) group with the baseline characteristics of the control group. We selected the following factors as covariates for PSM: sex, age, BMI, CCI score, ASA grade, history of abdominal surgery, preoperative NRS score, TNM, preoperative hemoglobin, preoperative albumin. Propensity scores were calculated by a binary logistic regression model using a 1:1 matching of patients using a neighborhood matching algorithm (a caliper size of 0.05). After matching, 302 patients were included in the study and the data was analyzed by univariate and multivariate analysis.

Statistical analysis

Continuous data were tested for normal distribution using the Kolmogorov-Smirnov method. Continuous data that conformed to a normal distribution are represented by their mean and standard deviation. Continuous data that did not conform to a normal distribution are represented using the median and quarterback distance [20]. Comparisons between normally distributed continuous variables were performed using the chi-square test or Fisher’s exact test. Conversely, comparisons between non-normally distributed continuous variables used the Mann-Whitney U test.

LDH was analyzed by receiver operating characteristic curve and the truncated value of 215.5 U/L was obtained, which was higher than the cutoff value in the ELDH group. Using PSM, 788 patients were matched 1:1 at the LDH cutoff to balance potential covariables between the two groups. Clinically relevant parameters were assessed using univariate analysis to identify potential risk factors associated with outcomes. Variables with p values < 0.10 in univariate analysis were included in multivariate analysis. Differences at p < 0.05 were considered statistically significant. All statistical analyses were performed using SPSS software for Windows (IBM version 25.0; Armonk, USA) and PSM was analyzed using the R Software for Windows (version 4.2.1).

Results

Patient characteristics

As shown in Table 1, we included a total of 788 patients before matching. According to the ROC curve analysis, the critical value of LDH was 215.5 U/L, and two groups were obtained, the ELDH group (n = 151) and the Control group (n = 637). We first compared the clinical characteristics of the two matched groups (Table 1). Univariate analysis showed that BMI index (p = 0.039) and sex (p = 0.045) were significantly different between the two groups. There were no statistically significant differences between the two groups in other characteristics, including age, albumin level, hemoglobin level, CCI score, ASA grade, NRS score, history of abdominal surgery, and TNM stage. The statistical difference between the two groups disappears after PSM analysis. Through this method, we obtained the matched data shown in Table 1 for the ELDH group (n = 151) and the Control group (n = 151). In the matched population, the median age of the ELDH group was 66 years, and the proportion of males was 67.5 percent. The median age of the control variable group was 67 years, and the proportion of males was 58.3%. After matching, there was no significant difference between the two groups by univariate analysis, thus removing the bias to some extent.

Table 1. The comparison of patient clinical characteristics before and after propensity score matching.

Characteristic	Patients before PSM			Patients after PSM
	Control group (N = 637)	ELDH group (N = 151)	P	Control group (N = 151)	ELDH group (N = 151)	P
Gender, n (%)			0.045^‡			0.095
Female	156	49		63	49
Male	481	102		88	102
Age, median (IQR), years *	65(29-85)	66(33-87)	0.143	67(32-85)	66(33-87)	0.815
BMI, median (IQR), kg/cm2 *	22.5(14.45-34.70)	22.66(15.57-32.01)	0.039^‡	22.48(14.45-31.99)	22.66(15.57-32.01)	0.275
Preoperative albumin, median (IQR), g/L*	38.5(16.0-49.9)	39.6(18.3-48.0)	0.750	38.8(16-45.5)	39.6(18.3-48.0)	0.456
Preoperative hemoglobin, median (IQR), g/L*	124(33-169)	129(56-168)	0.126	126(33-158)	129(56-168)	0.610
CCI, n (%)			0.724			0.458
0	392	88		98	88
1-2	153	38		30	38
3-6	92	25		23	25
ASA score, n (%)			0.443			0.165
1-2	555	128		136	128
3-4	82	23		15	23
NRS score, n (%)			0.544			0.725
0-2	405	92		89	92
3-6	232	59		62	59
History of abdominal surgery			0.617			0.749
Yes	83	22		24	22
No	554	129		127	129
T stage, n (%)			0.544			0.465
Tis/T1	172	40		41	40
T2	90	15		23	15
T3	120	32		33	32
T4	255	64		54	64
N stage, n (%)			0.506			0.924
N0	302	64		66	64
N1	104	23		26	23
N2	114	29		28	29
N3	117	35		31	35
TNM stages, n (%)			0.274			0.080
0/I	225	48		46	48
II	141	28		44	28
III	271	75		61	75

Abbreviations: BMI = body mass index, Charlson comorbidity index = CCI, ASA = American Society of Anesthesiologists, NRS = nutritional risk screening, TNM = tumor-node-metastasis.

Notes: Values are number of patients and percent unless indicated otherwise; * Values are median (inter quartile range). ‡ Values are statistically significant (P < 0.05);.

Situations related to readmission

Before matching, the overall 30-day readmission rate for the 788 patients was 7.2 percent, compared with 13.9 percent in the ELDH group and 5.7 percent in the control group. In the matched population, readmissions were counted, and as shown in Table 2, we demonstrate that 28 of the 302 patients in the analysis were readmitted within 30 days. Most of the readmissions were due to infection-related symptoms (n = 12, 45.9%), including intra-abdominal infection, incision infection, fever, and anastomotic leakage. The second most common readmission cause was gastrointestinal symptoms (n = 9, 32.1%), such as intestinal obstruction, nausea and vomiting, and gastroparesis.

Table 2. Reasons for readmission of gastric cancer patients.

Causes of Readmission	Total(N = 302)	Control group(N = 151)	ELD group(N = 151)
Infection-related Symptoms	12(3.97)	3(1.99)	9((5.96)
Intra-abdominal infection	5(1.66)	1(0.66)	4(2.65)
Incision infection	3(0.99)	1(0.66)	2(1.32)
Fever	2(0.66)	1(0.66)	1(0.66)
Anastomotic leakage	2(0.66)	0(0.00)	2(1.32)
Gastrointestinal Symptoms	9(2.98)	3(1.99)	6(3.97)
Intestinal obstruction	5(1.66)	1(0.66)	4(2.65)
Nausea and vomiting	3(0.99)	1(0.66)	2(1.32)
Gastroparesis	1(0.33)	1(0.66)	0(0.00)
Gastrointestinal bleeding	3(0.99)	1(0.66)	2(1.32)
Venous Thromboembolism	1(0.33)	0(0.00)	1(0.66)
Cerebral Infarction	1(0.33)	0(0.00)	1(0.66)
Liver Dysfunction	1(0.33)	0(0.00)	1(0.66)

Notes: Values are number of each readmission reason and percent unless indicated otherwise.

Univariate and multivariate analyses of risk factors for 30-day readmission

We performed univariate and multivariate analyses of readmissions in matched groups. Factors with p < 0.10 in univariate analysis were included in multivariate analysis. Univariate analysis showed between-group differences in LDH level (p = 0.005), ASA score (p = 0.007), and postoperative complications (p = 0.001) (Table 3). Moreover, multivariate logistic regression analysis adjusted for potential confounders showed that LDH (p = 0.005, odds ratio (OR) 3.768, 95% confidence interval (CI) 1.493–9.510), ASA score (p = 0.035, OR 2.784, 95%CI 1.074–7.214), and postoperative complication (p = 0.001, OR 4.156, 95%CI 1.802–9.583) were independent risk factors for 30-day readmission. In terms of expenditures, we compared the readmission group with the non-readmission group and found a statistical difference between the two groups (p = 0.010, OR 2.904, 95%CI 1.257–6.711).

Table 3. Univariate and multivariate analysis associated with 30-day readmission.

Risk factor	Univariate			Multivariate
	Non-readmitted	Readmitted	p Value	Odds ratio	95% CI		p Value
LDH			0.005^+	3.768	1.493-9.510		0.005^‡
Control group	144	7
ELDH group	130	21
Age, years			0.781
< 70	164	16
≥ 70	110	12
Gender			0.327
Female	104	8
Male	170	20
BMI, kg/cm2			0.818
< 25	210	22
≥ 25	64	6
Hypoalbuminemia			0.572
No	223	24
Yes	51	4
Anemia			0.373
No	198	18
Yes	76	10
ASA score			0.007^+	2.784	1.074-7.214		0.035^‡
1-2	244	20
3-4	30	8
CCI			0.705
0	168	18
1-2	61	7
3-6	45	3
NRS score			0.929
0-2	164	17
3-6	110	11
History of abdominal surgery			0.897
Yes	41	5
No	233	23
Sarcopenia			0.902
Yes	56	6
No	218	22
Time of Operation			0.928
< 210 min	159	16
≥ 210 min	115	12
Operation method			0.977
Laparoscopy	195	20
Open	79	8
Type of resection			0.570
Total gastrectomy	103	9
Subtotal gastrectomy	171	19
Methods of anastomosis			0.366
Billroth-I	122	11
Billroth-II	39	7
Roux-en-Y	113	10
T stage			0.624
Tis/T1	76	5
T2	33	5
T3	59	6
T4	106	12
N stage			0.841
N0	117	13
N1	46	3
N2	51	6
N3	60	6
TNM stages, n (%)			0.953
0/I	86	8
II	65	7
III	123	13
Postoperative Complication Grade			0.001^+	4.156	1.802-9.583		0.001^‡
<2	215	14
≥2	59	14
Length of postoperative stay			0.566
<13	133	12
≥ 13	141	16

BMI = body mass index, ASA = American Society of Anesthesiologists, Charlson comorbidity index = CCI, NRS = nutritional risk screening, TNM = tumor-node-metastasis.

Notes: Values are number of patients and percent unless indicated otherwise; +Values were included in multivariate analysis;‡ Values are statistically significant (P < 0.05).

Subgroup analysis for distinct subsets of patients

We performed a subgroup analysis to explore potential heterogeneity, as shown in Figure 2. We found no statistically significant differences between ELDH and controls in subgroups that included older age, males, obesity, sarcopenia, and high ASA scores. There were no differences among the other subgroups.

Figure 2. Subgroup analysis was conducted to explore the different level of LDH for distinct populations. ELDH, Elevated Lactate Dehydrogenase; ASA, American Society of Anesthesiologists; CI, confidence interval; OR, odds ratio; BMI, body mass index.

Analysis of postoperative outcomes

After univariate and multivariate analyses, postoperative complications were also associated with readmission. Therefore, we compared the postoperative outcomes between the ELDH group and the control group, as shown in Table 4. Postoperative complications were not the same as readmissions, and the same patient may have had multiple postoperative complications. Postoperative complications occurred in 41 of 302 patients. In the post-surgery complication group, 26 had infection-related complications. There was a correlation between ELDH and incision infection in 17 cases (p = 0.015). The length of hospital stay was significantly different between the ELDH group and the control group (p = 0.017, OR 2.180, 95%CI 1.136–4.182), which meant that the postoperative hospital stay in the ELDH group was more than twice as long as the control group. There was no statistically significant difference in the cost of first admission between the two groups.

Table 4. Postoperative complications and short-term outcomes.

Factors	Control group(N = 151)^b	ELDH group (N = 151)^b	P
Total complications, (n%)^a	41(27.15)	32(21.19)	0.226
Infection complications	26(17.22)	18(11.92)	0.192
Incision infection	7(4.64)	0(0.00)	0.015*
Pulmonary infection	9(5.96)	9(5.96)	1.000
Intra-abdominal infection	10(6.62)	6(3.97)	0.304
Anastomotic leakage	4(2.65)	7(4.64)	0.357
Septic shock	2(1.32)	1(0.66)	1.000
Skin infection	1(0.66)	1(0.66)	1.000
Urinary tract infection.	2(1.32)	0(0.00)	0.498
Gastrointestinal complications	5(3.31)	7(4.64)	0.556
Gastroparesis	3(1.99)	5(3.31)	0.723
Intestinal obstruction	1(0.66)	0(0.00)	1.000
Anastomotic stenosis	0(0.00)	3(1.99)	0.248
Singultus	1(0.66)	1(0.66)	1.000
Seroperitoneum	1(0.66)	1(0.66)	1.000
Pleural effusion	9(5.96)	8(5.30)	0.803
Gastrointestinal bleeding	5(3.31)	3(1.99)	0.723
gout	1(0.66)	0(0.00)	1.000
Hematological Disease	1(0.66)	0(0.00)	1.000
Anemia	4(2.65)	2(1.32)	0.684
Hypoalbuminemia	2(1.32)	4(2.65)	0.684
Heart Disease	0(0.00)	3(1.99)	0.248
Hepatic and renal dysfunction	1(0.66)	2(1.32)	1.000
Mental illness	2(1.32)	1(0.66)	1.000
Pulmonary embolism	0(0.00)	2(1.32)	0.498
Venous Thrombus Embolism	3(1.99)	5(3.31)	0.723
Cerebral Infarction	0(0.00)	2(1.32)	0.498
Serious complications
Clavien-Dindo classification ≥ III	17	13	0.442
Postoperative hospital stays, median (IQR) (days)	13.0(11.0-18.0)	13.0(11.0-18.0)	0.017*
Costs, median (IQR) (yuan)	57199.52(50736.88-71625.22)	57392.26(49496.99-72702.49)	0.474
Readmission within 30 days, n (%)	7	21	0.005*

^a Some patients have more than one type of postoperative complication. The total number of postoperative complications is higher than the number of patients who suffer the postoperative complication.

^b Values are number of patients and percent unless indicated otherwise;.

* Statistically significant (P < 0.05); IQR interquartile range.

Nomogram for readmission

A score system for accurately predicting readmission was established by multivariate logic analysis of 302 patients to identify independent risk factors, as shown in Figure 3. The nomogram had a C-index of 0.75. The ROC analysis in Figure 4 shows that the combined score had a higher value, and the area under the curve (AUC) was 0.75.

Figure 3. Nomogram for the risk of readmission within 30 days of discharge for patients with gastric cancer. ASA, American Society of Anesthesiologists; BMI, body mass index;.

Figure 4. Receiver operating characteristic curve for the prediction model Area under the curve was 0.75 (95% confidence interval 0.64–0.86).

Discussion

Gastric cancer is treated by upper gastrointestinal tract surgery, a relatively complex procedure with a high mortality rate. In our pre-matching data, the readmission rate at our study center was 7.2%, which is consistent with our prior study [21], as well as a related readmission study in Asian gastric cancer patients, which ranged from 6.6% to 8.7% [5]. By comparison, we show that the ELDH group has a steep readmission rate, 13.9% in the group with ELDH and 5.7% in the control group, and a multivariate logistic analysis after PSM demonstrated that there is a statistical difference. In addition, related studies suggested that increased mortality is associated with the occurrence of readmission [22], thus requiring early intervention and readmission. However, in the post-PSM data, we found no statistical difference in mean hospitalization costs between the ELDH group and the control group (57392.26 yuan and 57199.52 yuan, respectively). Patients at high risk of readmission will be under greater financial stress than those at low risk due to the possibility of a second hospitalization, especially in today’s pandemic environment. Thus, we believe that the financial pressures will be higher for the ELDH group. A growing number of countries and hospitals now include readmissions in their hospital quality management indicators, as higher readmission rates imply higher financial support. In particular, the readmission rates in European and American countries are significantly higher than those in Asia and may be related to the differences in Western social and cultural habits and the treatment of gastric cancer [23, 24]. The Affordable Care Act of 2010 in the United States, for example, was designed to reduce the number of readmission procedures in order to reduce insurance-related costs [4]. Thus, we can also reduce the burden on social care by analyzing readmissions while intervening for patients at risk of readmission.

In this study, we found that although the LDH level has an impact on the overall hospitalization period (p = 0.017), the two groups did not have a significantly different postoperative hospitalization time of more than 13 days (p = 0.566). There was no difference in the average hospitalization costs, but we found that higher LDH levels were an independent risk factor for readmission within 30 days after discharge of gastric cancer surgery patients (OR 3.768, 95%CI 1.493–9.510). LDH, a product of neoplastic necrosis due to hypoxia, may be a sign of high tumor burden [25]. Correlation analysis showed that high LDH levels indicate aggressive tumor biology and are a metabolic biomarker and immune-monitoring prognostic biomarker [26]. The risk of readmission associated with high LDH levels may be related to the aggressive tumor growth and metastasis [27]. When the tumor needs anaerobic glycolysis to generate more energy, the serum LDH levels rise accordingly. In addition, lactate produced by anaerobic glycolysis can further lead to a pH decrease at the cellular level, activating pathways that stimulate tumor growth and invasion [28–30]. Therefore, when the preoperative LDH level is high, the patient’s tumor cells may be in an active growth and invasion state, resulting in a poor short-term prognosis. When these patients with unnoticed poor prognoses are discharged, the risk of readmissions increases. Regarding the short-term outcomes, patients with higher preoperative serum LDH levels were more likely to have incision infections and readmission events. As shown in Table 2, the most frequent reasons for readmission were gastrointestinal and infectious symptoms. LDH is closely associated with inflammation [26, 31], having even been proposed in some disciplines as an indicator of early organ damage or inflammation [32, 33]. When tissues become infected, LDH is released, especially in patients with lung infections. Therefore, gastrointestinal and associated infectious symptoms were more likely to occur in the ELDH group. Gastrointestinal symptoms may lead to diaphragmatic dysfunction, increasing the burden on respiratory muscles and oxygen consumption, especially in patients with intestinal obstruction [34]. When oxygen consumption increases, anaerobic respiration also increases, especially when the patient’s oxygen intake does not increase or improve due to their symptoms, and serum LDH also increases accordingly. Patients with high LDH levels before surgery may be in this inflammatory state, and when they are discharged home after surgery, the inflammatory response of the disease itself is not entirely relieved, resulting in the worsening of inflammatory symptoms after discharge and readmission events.

In addition, we conducted an analysis of the sub-group, as shown in Figure 2. We demonstrate that LDH and readmission were statistically significant in most sub-groups. However, in the groups of elderly, men, obesity, sarcopenia, and high ASA scores, LDH and readmission were not statistically significant. With increased age and obesity, human LDH levels gradually decrease [35]; thus, we suggest that this situation may be affected by LDH distribution. Of course, another possibility is that the amount of data decreased after sub-group analysis, which leads to bias.

In summary, our findings suggest that LDH has great potential to be an effective indicator for assessing unplanned readmissions, which clinicians can use to examine patients for readmission and address the underlying etiology of readmission.

Unlike the patients in internal medicine, surgery patients are facing planned surgical procedures and have known surgical risks, which makes them at a high risk of readmission [36]. This was confirmed by our research showing a positive association between complications and readmissions (OR 4.156, 95%CI 1.802–9.583). Patients with complications during their first hospital stay were more likely to have readmissions after discharge. However, the reasons for readmission were more dependent on symptoms after discharge [37]. After PSM, a total of 73 patients developed postoperative complications, but only 19.2% (n = 14) of patients with postoperative complications were readmitted, with few readmissions for the same symptoms. In a large study of readmissions in general surgery patients, this view was echoed by Merkow et al. (2015) [38]. Therefore, we should conduct early interventions before discharging patients with complications after surgery, thereby improving the transition process of inpatient patients to outpatient patients.

The ASA score in the study represented the patient’s preoperative baseline status. In this study, we examined the association between the ASA score and readmission. Patients with an ASA score ≥3 have a higher probability of unplanned readmission after discharge. In the study by AC Meyer et al. (2021) [39], they concluded that patients with higher ASA scores were also at higher risk for cardiovascular complications. Meanwhile, in the analysis of Ahmad et al. (2009), preoperative cardiac disease was independently associated with readmission [13]. Combining both viewpoints, we postulate it might be th`e association between the ASA score and heart disease that leads to the association between the ASA score and readmission. However, no relevant analysis of cardiovascular disease was performed in this study, so we believe that future research needs examine this possibility.

Readmission will bring economic, physical, and psychological burdens to patients. Therefore, we obtained these independent risk factors for readmission through PSM analysis and established a scoring model to evaluate the risk probability of readmission. In Figure 3, we assigned the risk factors numerical values, multiplied each value by a coefficient, and added them together to calculate the probability. By ROC analysis, we found that the model area under the curve (AUC) of 0.78 was significantly larger than the AUC of a single independent factor. These data are easily available for patients. Therefore, we believe that this model can be used to evaluate the 30-day readmission probability of gastric cancer patients. Through our promoter diagram, the accurate probability of readmission for patients with gastric cancer patients can be provided to surgeons. The Nuo model is composed of the patient’s basic characteristics and related clinical features, which can be used after surgery. When the patients with a high risk of readmission are recognized using the promoter diagram, we can reduce the patient’s readmission likelihood by strengthening the relevant treatment and hospitalization.

TNM stage was not found to be a risk factor in gastric cancer studies [21], as we found in our study. A relationship between the TNM stage and readmission was found in readmission studies related to colorectal tumors [40]. The novelty of this study is that LDH is closely related to unplanned readmission after one month in gastric cancer patients through PSM and has great potential to become an effective indicator for evaluating unplanned readmission within one month for gastric cancer patients.

This study has some limitations. First, this is a single-center study, and there may be local bias, since when some patients develop symptoms after discharge, they may be directly hospitalized in local hospitals for reasons of travel or customs. Second, some patients were admitted due to a variety of complex factors. We chose the most obvious symptoms as the reason for readmission, which may lead to some reasons for readmission being ignored.

Conclusion

Patients with elevated preoperative LDH levels have a higher risk of readmission 30 days after surgery. LDH was an independent risk factor for 30-day postoperative readmission in gastric cancer patients, and it has the potential to become an evaluation index.

Author contributions

Wei-Sheng Chen: analyzed and interpreted the data, drafted and revised the manuscript; Ze-Xin Huang and Hui-Hui Zhang: acquired data; Yi-Qi Cai: carried out the statistical analysis; Xiao-Dong Chen: analyzed and revised the manuscript; Wen-Jing Chen: interpreted the data and revised the manuscript; Yun-Shi Huang and Guan-Bao Zhu: contributed toward the conception of the study, designed the study, analyzed and interpreted the data. All the authors read and approved the final manuscript.

Ethical Standards

The ethics committee of The First Affiliated Hospital of Wenzhou Medical University approved this study (approval number: 2014 No. 063) and the need for individual informed consent was waived.

Abbreviations
LDH	=	Lactate dehydrogenase
BMI	=	body mass index
ASA	=	American Society of Anesthesiologists
NRS	=	nutritional risk screening
TNM	=	tumor node metastasis
SD	=	standard deviations
IQR	=	interquartile range
PSM	=	propensity score matching

Acknowledgments

The authors would like to thank Editage (https://www.editage.com) for language editing support. There are no other commercial interests or sources of financial or material support to declare.

Disclosure statement

No potential conflict of interest was reported by the author(s).