Prognostic significance of preoperative haemoglobin A1c level in patients with lung adenocarcinoma

Abstract

OBJECTIVES

We investigated the influence of the preoperative haemoglobin A1c (HbA1c) value on the prognosis and pathology of patients with lung adenocarcinoma who underwent surgery.

METHODS

We reviewed the medical records of 400 lung adenocarcinoma patients who underwent lobectomy with mediastinal lymph node dissection between 2009 and 2013 using a prospectively maintained database. We stratified 400 patients into 4 groups according to the preoperative HbA1c value as follows: HbA1c ≤ 5.9 (n = 296), 6.0 ≤ HbA1c ≤ 6.9 (n = 70), 7.0 ≤ HbA1c ≤ 7.9 (n = 21) and HbA1c ≥ 8.0 (n = 12). We compared the recurrence-free survival and overall survival (OS) among these 4 groups. Univariate and multivariate analyses were performed to identify the risk factors for recurrence.

RESULTS

The median follow-up period was 61.2 months. On comparing the recurrence-free survival and OS rates among these 4 groups, we found that these rates among patients in the HbA1c ≥ 8.0 group were significantly poorer compared with the other 3 groups (5-year recurrence-free survival: HbA1c ≤ 5.9, 70.4%; 6.0 ≤ HbA1c ≤ 6.9, 69.7%; 7.0 ≤ HbA1c ≤ 7.9, 70.7%; ≥8.0 HbA1c, 18.8%; P = 0.002; and 5-year OS: HbA1c ≤ 5.9, 88.7%; 6.0 ≤ HbA1c ≤ 6.9, 80.6%; 7.0 ≤ HbA1c ≤ 7.9, 90.2%; ≥8.0 HbA1c, 66.7%; P = 0.046). Patients in the HbA1c ≥ 8.0 group had significantly more tumours with vascular invasion (P = 0.041) and experienced distant metastasis significantly more often (P = 0.028) than those with other values. A multivariate analysis revealed that preoperative HbA1c ≥ 8.0 [hazard ratio (HR) 2.33; P = 0.026] and lymph node metastasis (HR 3.94; P < 0.001) were significant independent prognostic factors for recurrence.

CONCLUSIONS

Our results revealed that preoperative HbA1c ≥ 8.0 is associated to poor prognosis due to the occurrence of distant metastasis and we should carefully follow these patients after surgery.

Clinical registration number

Hyogo Cancer Center, G-57.

Lung cancer is a major cause of cancer-related death worldwide [1].

INTRODUCTION

Lung cancer is a major cause of cancer-related death worldwide [1]. Adenocarcinoma is the most common histological subtype of lung cancer, and the proportion of adenocarcinoma cases is rising in many countries [2]. Diabetes mellitus (DM) is the most prevalent endocrine disorder, and its incidence is also increasing, significantly increasing the risk of mortality and morbidity worldwide [3]. The association between DM and cancer has recently garnered substantial attention.

Warburg proposed that high glycolytic activity might be related to the cancerous state, a phenomenon known as the ‘Warburg effect’ [4, 5]. The preferential utilization of anaerobic glycolysis for energy metabolism was observed in cancer cells, although this glycolytic pathway is much less efficient for adenosine triphosphate (ATP) production than the mitochondrial oxidative phosphorylation preferentially utilized in other normal cells [5, 6]. As such, this high-glucose condition is preferable for cancer cell proliferation, as cancer cells need more glucose to produce energy [5]. Indeed, several researchers have reported that DM has a negative influence on the prognosis of lung cancer patients [7–12]. However, thus far, there have been few studies reporting on the association between DM and the prognosis of patients with lung cancer who underwent surgical resection, with the findings obtained being controversial [13–16]. Lung cancer has been reported to show histology-specific glucose metabolism [17, 18]. A recent study showed that tumours with an increased glucose uptake were functionally enriched for molecular process associated with invasion in lung adenocarcinoma and cell growth in squamous cell carcinoma [18]. Considering these metabolic differences, study designs including patients with all histologies might have obscured the influence of DM on lung cancer.

In the present study, we analysed DM and non-DM lung adenocarcinoma patients who underwent lobectomy with mediastinal lymph node dissection and investigated the influence of DM on the prognosis and pathology of these patients. We used the preoperative haemoglobin (Hb) A1c value as an indicator of the severity of DM.

PATIENTS AND METHODS

Ethical statement

The Hyogo Cancer Center Institutional Review Boards approved the study (No. G-57), and informed consent was obtained from all patients.

Patient cohort

We retrospectively reviewed a prospectively maintained clinical database of 1091 lung cancer patients who underwent surgical resection at Hyogo Cancer Center from 1 January 2009 to 31 December 2013. Among them, we excluded 691 patients for the following reasons: histological subtypes other than adenocarcinoma (n = 345), performed sublobar resection or lobectomy without mediastinal lymph node dissection (n = 279), incomplete resection or pathological stage 4 (n = 13), synchronous multiple lung cancers (n = 28), performing preoperative treatment (n = 15) and insufficient medical records (n = 11). The 400 remaining patients were analysed in this study (Fig. 1).

Figure 1:

The patient population in this study. HbA1c: haemoglobin A1c.

We collected information on the patients’ age, gender, smoking history, lung function, tumour marker, anti-diabetic treatment, surgical approach, postoperative complication, adjuvant chemotherapy and pathological data. The disease stage was determined based on the seventh edition of tumour, node and metastasis (TNM) classification using the International Union Against Cancer (UICC) staging system [19]. Surgery was performed for patients with an Eastern Cooperative Oncology Group Performance Status Scale of 0 or 1 [20].

The preoperative examination and follow-up

Contrast-enhanced chest and abdominal computed tomography (CT), positron emission tomography–CT and brain magnetic resonance imaging (MRI) were performed for preoperative staging. All patients received a blood test before surgery and had their haemoglobin A1c (HbA1c) level measured by the Japan Diabetes Society (JDS) method [21]. Most of the patients had their HbA1c level measured within 1 month before surgery.

After surgery, patients with stage IB disease were suggested to take uracil and tegafur orally for 2 years and patients with stage II or III disease were suggested to receive 4 cycles of intravenous chemotherapy. Patients were evaluated postoperatively at 3-month intervals for 2 years, at 6-month intervals for the subsequent 3 years and annually thereafter. Follow-up examinations included chest radiography, contrast-enhanced CT, brain MRI and bone scintigraphy as well as haematological and biochemical analyses, including the measurement of the tumour markers. The recurrence-free survival (RFS) and overall survival (OS) were calculated according to the Kaplan–Meier method, and the log-rank test was used to evaluate differences in the distributions. The RFS was defined as the time interval between the date of surgery and the date of death without recurrence or the date of the first recurrence detected by a radiological examination. The OS was defined as the time interval between the date of surgery and the date of death. Local recurrence was defined as any recurrence within the same lung, ipsilateral lymph nodes or pulmonary hilum. Distant recurrence was defined as any recurrence other than local recurrence.

Statistical analyses

Statistical analyses were performed using the JMP 14 software program (SAS Institute, Cary, NC, USA). The RFS and OS were calculated according to the Kaplan–Meier method, and differences in the distributions were evaluated by the log-rank test. Student’s t-test and the χ² test were performed to assess the significance of the differences in age, gender, smoking history, lung function, tumour marker, pathological stage (p-stage), adjuvant chemotherapy, recurrence site and other pathological factors between the HbA1c ≤ 7.9 and HbA1c ≥ 8.0 groups. A Cox proportional hazards model was used to evaluate the relationship between the clinicopathological factors and RFS rate after surgery with the hazard ratio (HR) and 95% confidence interval (CI). P-values <0.050 were considered to indicate statistical significance. P-values may not be interpreted as confirmatory but rather descriptive.

RESULTS

Comparing the recurrence-free survival and overall survival among groups stratified by HbA1c values

The median follow-up period was 61.2 months. There was no operative death within 30 days postoperatively in this patient cohort. We analysed whether or not the preoperative HbA1c value affected the prognosis of patients with lung adenocarcinoma after surgery. We stratified 400 patients into 4 groups according to HbA1c value as follows: HbA1c ≤ 5.9 (n = 296), 6.0 ≤ HbA1c ≤ 6.9 (n = 70), 7.0 ≤ HbA1c ≤ 7.9 (n = 21) and ≥8.0 HbA1c (n = 12). On comparing the RFS and OS rates among these 4 groups, we found that these rates among patients in the 8.0 ≤ HbA1c group were significantly poorer compared with the other 3 groups (5-year RFS: HbA1c ≤ 5.9, 70.4%; 6.0 ≤ HbA1c ≤ 6.9, 69.7%; 7.0 ≤ HbA1c ≤ 7.9, 70.7%; ≥8.0 HbA1c, 18.8%; P = 0.002 and 5-year OS: HbA1c ≤ 5.9, 81.2%; 6.0 ≤ HbA1c ≤ 6.9, 71.3%; 7.0 ≤ HbA1c ≤ 7.9, 79.9%; ≥8.0 HbA1c, 57.1%; P = 0.046) (Fig. 2). This tendency was also observed in pathological stage 1 patients (5-year RFS: HbA1c ≤ 5.9, 87.0%; 6.0 ≤ HbA1c ≤ 6.9, 81.5%; 7.0 ≤ HbA1c ≤ 7.9, 85.1%; ≥8.0 HbA1c, 0%; P < 0.01 and 5-year OS: HbA1c ≤ 5.9, 84.4%; 6.0 ≤ HbA1c ≤ 6.9, 81.2%; 7.0 ≤ HbA1c ≤ 7.9, 84.6%; ≥8.0 HbA1c, 53.6%; P < 0.001) (Fig. 3).

Figure 2:

The comparison of the recurrence-free survival (A) and overall survival (B) among all-stage patients stratified by the value of HbA1c. HbA1c: haemoglobin A1c.

Figure 3:

The comparison of the recurrence-free survival (A) and overall survival (B) among pathological stage 1 patients stratified by the value of HbA1c. HbA1c: haemoglobin A1c.

Comparing the clinicopathological characteristics and first recurrence site between patients in the HbA1c ≤ 7.9 and ≥8.0 HbA1c groups

Because the prognosis of patients in the ≥8.0 HbA1c group was significantly poorer than that in other groups, we focused on the patient groups and continued to analyse the clinicopathological differences between patients in the HbA1c ≤ 7.9 and ≥8.0 HbA1c groups. The results are summarized in Table 1 and Supplementary Table 1. There was no significant difference between the groups in the age, gender, smoking history, carcinoembryonic antigen, pathological stage, tumour size, lymph node metastasis, pleural invasion, lymphatic permeation, pulmonary metastasis, surgical approach, postoperative complication or adjuvant chemotherapy. However, the ≥8.0HbA1c group included significantly more patients with vascular invasion (P = 0.042).

Table 1:

The clinicopathological characteristics of the patients

Factors	HbA1c ≤ 7.9 (n = 388)	HbA1c ≥ 8.0 (n = 12)	P-value
Age (years), mean (range)	67.2 (35–84)	67.0 (54–80)	0.954
Gender, n (%)
Male	216 (56)	6 (50)	0.697
Female	172 (44)	6 (50)
Smoking history, n (%)
Former or current	204 (53)	7 (58)	0.694
Never smokers	184 (47)	5 (42)
Lung function
FEV1.0%, mean	74.4	73.3	0.679
Tumour marker
CEA, mean	7.4	5.7	0.375
Anti-diabetic treatment, n (%)
None	353 (91)	2 (17)
Oral medicine only	30 (8)	5 (42)	<0.001
Use of insulin	5 (1)	5 (42)
Stage, n (%)
IA	170 (44)	3 (25)	0.334
IB	87 (22)	4 (33)
IIA	43 (11)	3 (25)
IIB	17 (4)	1 (8)
IIIA	71 (18)	1 (8)
Tumour size (mm), mean (range)	30.6 (3–140)	37.5 (12–75)	0.236
Lymph node metastasis, n (%)
pN 0	287 (74)	10 (83)	0.728
pN 1	37 (10)	1 (8)
pN 2	64 (16)	1 (8)
Pleural invasion, n (%)
(−)	274 (71)	6 (50)	0.124
(+)	114 (29)	6 (50)
Vascular invasion, n (%)
(−)	242 (62)	4 (33)	0.041
(+)	146 (38)	8 (67)
Lymphatic permeation, n (%)
(−)	273 (70)	6 (50)	0.130
(+)	115 (30)	6 (50)
Pulmonary metastasis in the same lobe, n (%)
(−)	375 (97)	12 (100)	0.812
(+)	13 (3)	0 (0)
Surgical approach, n (%)
Thoracotomy	274 (71)	9 (75)	0.742
VATS	114 (29)	3 (3)
Postoperative complication, n (%)
None	319 (82)	9 (75)	0.208
Minor	63 (16)	2 (17)
Major	6 (2)	1 (8)
Adjuvant chemotherapy, n (%)
Performed	176 (45)	4 (33)	0.343

CEA: carcinoembryonic antigen; FEV1.0%: forced expiratory volume in 1 s; HbA1c: haemoglobin A1c; VATS: video-assisted thoracoscopic surgery.

Factors related to the recurrence-free survival

We performed univariate and multivariate analyses to identify the factors related to the RFS, and the results are shown in Table 2. Univariate and multivariate analyses were performed using 9 clinical parameters [age, gender, smoking history, lung function, preoperative HbA1c ≥8.0, carcinoembryonic antigen, adjuvant chemotherapy, tumour size (>30 mm)]. A multivariate analysis revealed that preoperative HbA1c ≥8.0 (HR 2.33; P = 0.026) and lymph node metastasis (HR 3.94; P < 0.001) were significant independent prognostic factors for recurrence.

Table 2:

Univariate and multivariate analyses for the recurrence-free survival

	Univariate analysis			Multivariate analysis
Variables	HR	95% CI	P-value	HR	95% CI	P-value
Age (≥70 years)	1.40	1.06–1.84	0.020	1.01	0.99–1.04	0.224
Gender (male)	1.25	0.95–1.65	0.112	1.27	0.75–2.15	0.361
Smoking history (ever smoker)	1.30	0.98–1.71	0.067	0.77	0.45–1.31	0.333
FEV1.0% (<70%)	1.26	0.91–1.71	0.163	1.01	0.66–1.50	0.977
HbA1c (≥8.0)	2.45	1.10–4.65	0.034	2.33	1.07–4.48	0.026
CEA (≥5.0)	2.28	1.72–3.01	0.005	1.05	0.73–1.50	0.813
Chemotherapy (performed)	1.61	1.21–2.14	0.009	0.71	0.47–1.05	0.094
Tumour size (≥30)	5.14	2.32–10.3	0.007	1.40	0.99–1.98	0.062
N factor (≥1)	7.71	5.85–10.2	<0.001	3.94	2.54–6.12	<0.001

CEA: carcinoembryonic antigen; CI: confidence interval; FEV1.0%: forced expiratory volume in 1 s; HbA1c: haemoglobin A1c; HR: hazard ratio.

Comparing the first recurrence site

We compared the first recurrence site between patients in the HbA1c ≤ 7.9 and ≥8.0 HbA1c groups, and the results are shown in Table 3. The rate of distant recurrence was significantly higher in patients with ≥8.0 HbA1c than in those with HbA1c ≤ 7.9 (P = 0.028). The first recurrence sites in patients who experienced distant recurrence only were brain metastasis (n = 2), contralateral lung metastasis (n = 1), adrenal gland (n = 1), liver (n = 1) and small intestine (n = 1).

Table 3:

The comparison of the first recurrence site

Recurrence	HbA1c < 8.0 (n = 125), n (%)	HbA1c ≥ 8.0 (n = 8) n (%)	P = 0.028
Local + distant	37 (30)	1 (12.5)
Distant only	37 (30)	6 (75)
Local only	51 (40)	1 (12.5)

HbA1c: haemoglobin A1c.

DISCUSSION

In the present study, we demonstrated that lung adenocarcinoma in patients with preoperative HbA1c ≥ 8.0 metastasized significantly more frequently and showed a significantly poorer prognosis after surgery than in patients with preoperative HbA1c ≤ 7.9. Multivariate analyses revealed that preoperative HbA1c ≥ 8.0 as well as lymph node metastasis were independent prognostic factors for recurrence.

Many researchers have reported the association between cancer and DM. Rao Kondapally Seshasai et al. [11] analysed 829 900 people in 97 prospective studies and found that DM is associated with substantial premature death from several cancers. In that study, fasting glucose levels exceeding 100 mg/dl (5.6 mmol/l) were associated with cancer death [11]. However, results regarding the association between the prognosis of lung cancer and DM have been conflicting. Hatlen et al. [7] analysed data from 1677 patients in 4 studies and reported that patients with lung cancer with DM had an increased survival compared with those without DM. In contrast, Zhu et al. [15] reviewed 12 articles with a total of 15 180 lung cancer patients and reported an association between DM and an inferior prognosis among lung cancer patients. They found that DM had a more negative influence on the OS of surgically treated patients (HR 1.70, 95% CI 0.94–3.08) than on that of non-surgically treated patients. However, Dhillon et al. [16] reported that DM had no influence on the OS in pathological stage I non-small lung cancer (NSCLC) patients undergoing surgical resection.

Few reports have described the relationship between DM and the prognosis in patients with NSCLC who underwent surgical resection, and these results have been controversial [13, 14, 22]. As these previous studies included patients with all histologies of NSCLC and who received sublobar resection, it might have obscured the actual influence of DM on lung cancer. Lung cancer was reported to have histology-specific glucose metabolism, and lung adenocarcinoma has its own specific type of metabolism [4–6, 23]. Therefore, we considered it best to include only patients with lung adenocarcinoma in our present study. In addition, patients with severe DM were less likely to receive aggressive treatment than others, due to concerns about complications. Given these issues, we included only lung adenocarcinoma patients who had undergone lobectomy with mediastinal lymph node dissection in our study in order to reduce selection biases.

In the present study, we used the preoperative HbA1c value as an indicator of the severity of DM. The HbA1c value reflects the mean blood glucose levels over the last 8–12 weeks. A preoperative HbA1c value of ≥8.0 indicated poorly a controlled preoperative blood sugar state [23]. This high-glucose condition and state of metabolic disorder in cancer cells are considered to accelerate cancer cell proliferation by the mechanisms described in the Introduction section [5]. In addition, it was previously reported that there is a relationship between the epidermal growth factor (EGF) pathway and cancer metabolism [24, 25]. The high-glucose condition is considered to promote cancer cell proliferation via the induction of EGF expression and transactivation of EGF receptor [24]. Elevated insulin levels in patients with severe DM are also considered to affect cancer progression through the insulin-like growth factor-1 pathway [26]. Insulin-like growth factor-1 pathway is reportedly associated with tumourigenesis [27] and tumour progression [28, 29]. In the present study, the tumours in patients with HbA1c ≥ 8.0 were more aggressive and often caused vascular invasion and distant metastasis than those in other patients, and these proposed biological mechanisms might be associated with our results.

Limitation

Several limitations associated with the present study warrant mention. First, this study was a retrospective, non-randomized, single-centre study. The number of patients with preoperative HbA1c ≥ 8.0 was small (n = 12) compared with the other patient cohorts. Second, information on histological subtype of adenocarcinoma and genetic abbreviations was not included in this study. Genetic abbreviations, such as an EGF receptor mutation or anaplastic lymphoma kinase rearrangement, might have affected the outcomes after surgery. Third, we only evaluated the preoperative HbA1c value, which represented preoperative glycaemic control; therefore, we were unable to evaluate the therapy course of DM after surgery, such as whether patients had been treated with dietary therapy or had their condition controlled by taking medications (e.g. insulin or oral agents). Fourth, HbA1c value was measured by the JDS method in this study; thus, this value needs to be adjusted by adding 0.4% to the HbA1c value (JDS) to apply to the current international measurement method of the National Glycohaemoglobin Standardization Program (NGSP) [21].

However, despite our study’s limitations, we feel that our results are significant. Further studies will be needed to verify the validity of our findings.

CONCLUSIONS

The present study showed that preoperative HbA1c ≥ 8.0 was a prognostic factor for patients who underwent surgery for lung adenocarcinoma. As patients with preoperative HbA1c ≥ 8.0 often experience distant metastasis, we should carefully follow these patients after surgery. With regard to diabetic control, it might be better to be maintain an HbA1c value of <8.0% (JDS, NGSP: 8.4%) before surgery.

SUPPLEMENTARY MATERIAL

Supplementary material is available at ICVTS online.

Conflict of interest: none declared.

Author contributions

Hiroyuki Ogawa: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Validation; Visualization; Writing—original draft; Writing—review & editing. Yusuke Fujibayashi: Data curation; Supervision. Megumi Nishikubo: Data curation; Supervision. Yuki Nishioka: Data curation; Supervision. Shinya Tane: Data curation; Supervision; Validation. Yoshitaka Kitamura: Data curation; Supervision; Validation. Wataru Nishio: Supervision; Validation.

Reviewer information

Interactive CardioVascular and Thoracic Surgery thanks David G. Healy, Akif Turna, Gonzalo Varela and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

ABBREVIATIONS

CI

Confidence interval

CT

Computed tomography

DM

Diabetes mellitus

EGF

Epidermal growth factor

HbA1c

Haemoglobin A1c

HR

Hazard ratio

JDS

Japan Diabetes Society

MRI

Magnetic resonance imaging

NSCLC

Non-small-cell lung cancer

OS

Overall survival

RFS

Recurrence-free survival