This article is mentioned in:

Introduction

Lung cancer (LC) is one of the leading causes of cancer-related morbidity and mortality worldwide (1,2). In 2020, ~2.2 million new cases and 1.8 million deaths associated with LC were reported, accounting for ~11% of all cancer diagnoses and 18% of cancer deaths (3). In addition, data from 2017 indicate that LC was responsible for 40 million disability-adjusted life years (4). It has a notable prevalence among the elderly population, with an average age of diagnosis of ~70 years (5).

With recent advances in diagnostic techniques, treatment modalities and personalized medicine, there are more efforts to better understand several factors that may impact the course and progression of LC (6–8). Numerous studies are focusing on the interplay between systemic health conditions, such as renal dysfunction, and cancer outcomes. Renal dysfunction, which may range from a mild decline in glomerular filtration rate (GFR) to end-stage renal disease (9), affects millions of individuals globally (10). Its intricate relationship with cancer outcomes has emerged as an area of growing interest in oncology (9–11). Kidneys can potentially influence cancer progression and treatment response through mechanisms that may include altered drug metabolism, immune system modulation, inflammation and disruption of hormonal pathways (11–13). Renal insufficiency (RI) is particularly prevalent in elderly patients, where there is an annual decline of ~1% in the average creatinine clearance (14).

In several cancer types, the diagnosis of chronic kidney disease (CKD) as a comorbidity has been associated with unfavorable prognoses (15). However, the exact impact of CKD on the prognosis of patients with LC remains a subject of debate. Whilst certain studies have reported no significant association between CKD and overall survival (OS) and/or disease-free survival (DFS) in patients with LC, others report different outcomes (16–18). These discrepancies could be attributed to differences in study populations, methodologies and outcome measures. To date, no systematic reviews and meta-analyses on the influence of RI and/or CKD on LC outcomes have been performed, to the best of our knowledge. Therefore, the present study aimed to systematically assess and quantify the association between renal dysfunction and/or CKD and the survival outcomes (OS and DFS) of LC.

Materials and methods

Search for relevant literature

An extensive systematic literature search of the PubMed (https://pubmed.ncbi.nlm.nih.gov/), Embase (https://www.embase.com/landing?status=grey) and Scopus (https://www.scopus.com/home.uri) databases was performed. The search method of the present study incorporated relevant keywords, synonyms and Medical Subject Headings (MeSH) terms. Furthermore, search terms that allowed keyword placement specification were used. In PubMed, the (tiab) tag was used to search for key words in titles and abstracts. The search strategies used in the PubMed, Embase and Scopus databases are presented in Table SI. Notably, the term ‘renal injury’ was not included in the search strategies as it tends to refer to acute conditions, such as acute kidney injury, which are different from chronic conditions, such as CKD, that the present study intended to focus on. Moreover, the present study aimed to assess the intersection of CKD and LC, specifically looking at long-term clinical outcomes, and using specific search terms ensured that the studies relevant to chronic conditions were retrieved, rather than acute episodes of kidney damage, which have different etiologies, treatment approaches and outcomes.

The searches were limited to human studies and articles published in the English language. Furthermore, the search scope was confined to studies published from the inception of the databases up to July 31st, 2023. Reference lists and relevant review articles were also manually reviewed to complement the electronic search and ensure the inclusion of any potentially overlooked studies.

The study inclusion criteria were as follows: i) Observational studies (cohort and case-control) and clinical trials that evaluated the association between RI/CKD and OS and/or DFS in patients with LC; ii) studies in adult human populations with confirmed LC diagnoses; iii) studies with the criteria for RI or CKD specified; iv) studies with a comparator group (normal renal function); and v) studies reporting relevant outcomes, such as OS, DFS and progression-free survival. The study exclusion criteria were as follows: i) Studies with overlapping or insufficient data; ii) studies that used inappropriate comparators i.e., comparison group does not comprise subjects with LC and normal kidney function, or the study did not have any comparison group; and iii) review articles, conference abstracts and case reports.

Screening, selection and data extraction

After performing the search strategy across the three databases and assembling the initial pool of studies, duplicate entries were removed. After an initial assessment of titles and abstracts, two independent study authors performed a comprehensive review of full-text articles, evaluating the eligibility of each study based on the predetermined criteria. In the event of differences in judgment, disagreements were resolved through discussion, and if necessary, the perspective of a third author was sought.

To enhance reliability throughout the article selection process, study authors involved in the screening and selecting articles underwent training sessions that included detailed guidelines and examples illustrating the application of inclusion and exclusion criteria. Calibration exercises were performed to ensure consistency in decision-making. Inter-rater reliability statistic, namely, Cohen's κ coefficient, was calculated to assess agreement between study authors (κ ranged from 0.88–0.97). Meetings were held regularly to discuss challenging cases and clarify ambiguous situations, ensuring all decisions were well-founded and transparently documented.

The present research adhered to the Preferred Reporting Items for Systematic reviews and Meta-Analyses guidelines to ensure the transparency and rigor of the methodology (19). The study protocol was formally registered in the International Prospective Register of Systematic Reviews (registration no. CRD42023455318). A total of two authors independently extracted data into a standardized form. The Newcastle-Ottawa Scale (NOS) assessed the quality of the included studies (20).

Statistical analysis

Statistical analyses were performed using STATA version 15.0 (StataCorp LP). As the outcomes were categorical and observed across extended follow-up periods, the combined effect size is presented as the hazard ratio (HR) and 95% confidence interval (CI). A random-effects model was used for all the analyses in the present study to account for variances in the baseline characteristics among the studies included. Group analyses were based on the stage and histological type of LC, the chosen treatment approach and the criteria for assessing RI/CKD. To evaluate potential publication bias, Egger's test and funnel plots were used (21). P<0.05 was considered to indicate a statistically significant difference.

Results

Flow of study selection

The search approach identified a total of 295 studies. After removing 81 duplicates, the remaining 214 studies underwent initial screening based on their titles and abstracts, and an additional 197 reports were removed. These reports were excluded because, based on the title and the abstract, the articles were found to be non-relevant to the scientific question being addressed. Full-text examination of the remaining 17 studies resulted in the exclusion of an additional seven studies. Out of these seven excluded studies, three studies did not have an appropriate control group, two studies had data that overlapped with studies already included for the meta-analysis and two were case reports. Ultimately, the meta-analysis included 10 studies (16–18,22–28). Fig. 1 illustrates the selection process of the studies in the present review.

Figure 1. Selection process of the studies included in the present review.

Characteristics of the included studies

Table I provides details of the included studies. All studies were observational in design and used retrospective data. Studies were performed in Japan (n=2), the United States (n=2), China (n=2), Taiwan (n=2), the Republic of Korea (n=1) and France (n=1). The duration of follow-up was 12 months to 5 years. In seven studies, most of the patients had non-small cell (NSC) LC. In four studies, most patients had stage I or II LC, and in five, patients had stage III or IV tumors. A total of four studies reported predominantly surgical management of LC, whilst in another 4 studies, medical management (such as chemotherapy, radiotherapy or targeted therapy) was used. A total of two studies did not report on the management modality. The diagnosis of RI or CKD was based on different criteria in the included studies: A total of three studies used International Statistical Classification of Diseases (ICD)-9-based diagnoses (29), and another three studies used the CKD-Epidemiology Collaboration (CKD-EPI) equation (30) and considered an estimated (e)GFR of <60 ml/min/1.73 m2 to denote the presence of RI/CKD (31). The remaining studies used definitions of RI/CKD based on serum creatinine, the Cockcroft-Gault formula and the Modification of Diet in Renal Disease equation (32,33). A total of five studies reported marked baseline differences in the prevalence of comorbidities among patients with and without RI/CKD. Patients with RI/CKD had a notably higher incidence of diabetes, hypertension, cardiovascular diseases and anemia. All included studies reported adjusted effect sizes using multivariable regression analysis. Factors such as age, sex, comorbidities and other relevant parameters differing between groups, were accounted for in these analyses. The total number of included patients in all 10 studies was 35,249 (4,302 patients with RI and 30,947 patients without RI). The mean quality assessment score of all studies was 7.1 (acceptable quality). A total of seven studies had a NOS score of 7 (with 9 being the maximum), two had a score of 8, and one study had a score of 6 (Table I).

Table I. Study details.

Table I.

Study details.

First author/s, year	Study design	Location	Age	Sex	Tumor type and stage	Mode of management	Follow-up duration	Definitions used	Sample size	Newcastle-Ottawa quality score	Adjustment/matching variables	(Refs.)
Lu et al, 2023	Retrospective observational	Taiwan	Median age of ~68 years	Male (64%)	NSCLC (78%); stage III–IV (70%)	Medical treatment (chemotherapy and/or radiotherapy or targeted therapy) in 40%	Not reported	End-stage renal disease defined as per ICD-9 code 585	CKD, n=133 and no CKD, n=532	7	Matching performed for age, sex, cancer histological type, cancer stage and treatment type	(16)
Saito et al, 2022	Retrospective observational	Japan	Median age of 69 years	Male (~60%)	Subjects had NSCLC with stage I–II	Surgery	~5 years	RI defined as serum creatinine ≥1.5 mg/dl	RI, n=113 and no RI, n=16,169	7	Age, sex, renal function status, smoking, comorbidities, pathological stage, histological type, grade and surgical procedure	(17)
Cho et al, 2021	Retrospective observational	Republic of Korea	Mean age of ~64 years	Male (~70%)	Subjects had NSCLC with stage III–IV	Surgery with or without chemotherapy (52%)	Mean follow up: 2.6 years	RI defined as eGFR <60 ml/min/1.73 m2; eGFR calculated based on CKD-EPI equation	RI, n=1,419 and no RI, n=1,783	7	Age, sex, BMI, smoking, initial eGFR, anemia, hyponatremia and pathological subtype	(22)
Jia et al, 2020	Retrospective observational	China	54% aged >60 years	Male (68%)	Subjects had NSCLC; stage I–II (86%)	Radical surgery	Not reported	CKD defined as eGFR <60 ml/min/1.73 m2; eGFR calculated based on CKD-EPI equation	CKD, n=63 and no CKD, n=77	7	Age, sex, smoking, drinking, TNM stage, tumor histology/differentiation/maximum tumor diameter, eGFR, adjuvant or neoadjuvant treatment, and anemia	(23)
Magali et al, 2020	Retrospective observational	France	Mean age of ~60 years	Male (~60%)	Subjects had NSCLC with stage III–IV	Medical treatment (chemotherapy)	12 months	RI defined as eGFR 60–89 ml/min/1.73 m2; eGFR calculated using the MDRD formula	RI, n=25 and no RI, n=87	7	Age, sex, GFR at diagnosis, cardiovascular disease, diabetes, malnutrition, and cumulated dose of cisplatin and pemetrexed	(24)
Yamamoto et al, 2019	Retrospective observational	Japan	Mean age of ~70 years	Male (58%)	Subjects had NSCLC; Majority with stage I–II	Surgical (lobectomy)	Median follow-up of ~50 months	CKD defined as eGFR <60 ml/min/1.73 m2; eGFR calculated based on Clinical Practice Guideline book for the diagnosis and treatment of CKD (2012) published by Japanese Society of Nephrology	CKD, n=55 and no CKD, n=616	8	Age, sex, smoking, hypertension, diabetes, COPD, histology, pathological stage, post-operative complication and surgical approach (VATS or thoracotomy)	(25)
Wei et al, 2018	Retrospective observational	Taiwan	Mean age of ~75 years early stage and	73% males	Early stage; histologic type of tumor not reported	~80% received non-surgical treatment	~5 years	Diagnosis of CKD was determined using specific ICD codes and one of the following criteria: i) ≥3 outpatient visits within a 6-month period, each with a diagnosis of CKD; or ii) inpatients with diagnosis of CKD on admission	CKD, n=2,269 and no CKD, n=9,076	7	Age, type of surgery and multiple comorbidities (hypertension, diabetes mellitus and COPD, congestive heart failure, cholelithiasis and hyperlipidemia)	(18)
Yang et al, 2016	Retrospective observational	China	Mean age of 53 years	Male (59%)	Histologic type of tumor not reported; Majority with stage III–IV	Details on mode of management not provided	Median follow-up of 40 months	RI defined as eGFR <60 ml/min/1.73 m2; eGFR calculated based on CKD-EPI equation	RI, n=56 and no CKD, n=1,323	8	Age, sex, eGFR, tumor stage, hypertension, diabetes, CVD, smoking and presence of proteinuria	(26)
Kutluk Cenik et al, 2013	Retrospective observational	United States	Mean age of ~58 years	Male (57%)	Subjects had NSCLC with stage III–IV	Medical treatment (chemotherapy)	Not provided	CKD defined as eGFR <60 ml/min/1.73 m2; eGFR calculated using the CKD-CG formula	CKD, n=101 and no CKD, n=197	7	Age, sex and ethnicity	(27)
Tammemagi et al, 2003	Retrospective observational	United States	Not reported	Male (59%)	Not reported	Not reported	Median follow up of 819 days	End-stage renal disease defined as per ICD-9	CKD, n=68 and no CKD, n=1,087	6	Age, sex, smoking status, histology and stage	(28)

[i] NSCLC, non-small cell lung cancer; ICD, International Statistical Classification of Diseases; RI, renal insufficiency; CKD, chronic kidney disease; CVD, cardiovascular disease; GFR, glomerular filtration rate; eGFR, estimated GFR; CKD-EPI, CKD epidemiology collaboration; BMI, body mass index; TNM, tumor-node-metastasis; MDRD, Modification of Diet in Renal Disease; COPD, chronic obstructive pulmonary disease; VATS, video-assisted thorascopic surgery; CKD-CG, CKD-Cockcroft-Gault.

Findings for the outcomes of interest

Patients with RI/CKD had a notably worse OS at follow-up (HR, 1.38; 95% CI, 1.16–1.63; n=10; I2=60.5%) compared with patients without RI/CKD (Fig. 2). There was no apparent publication bias demonstrated by Egger's test (P=0.98) and the funnel plot (Fig. S1). Subgroup analysis demonstrated a marked association between poor OS and RI/CKD in patients with stage I/II LC (HR, 1.76; 95% CI, 1.30–2.37; n=5; I2=58.6%), but not in patients with stage III/IV LC (HR, 1.18; 95% CI, 0.91, 1.54; n=6; I2=64.6%; Table II). Furthermore, when subgroup analysis was done based on methods to diagnose CKD, studies that used an ICD-9-based diagnosis reported a notable association between presence of RI/CKD and poor OS (HR, 1.40; 95% CI, 1.14–1.71; n=3; I2=52.0%), when compared to those with LC and no associated RI/CKD. No such association was detected in studies that used an CKD-EPI-based diagnosis (HR, 1.55; 95% CI, 0.97–2.47; n=3, I2=58.6%; Table II). Similarly, we conducted subgroup analysis based on the histologic type of LC. In this analysis, when only studies with patients with NSCLC were analyzed, a marked association between RI/CKD and poor OS was demonstrated (HR, 1.54; 95% CI, 1.22–1.94; n=7; I2=33.8%), compared to those with LC and no associated RI/CKD. Irrespective of the treatment modality, RI/CKD had a notable association with poor OS at follow-up (surgical management: HR, 1.78; 95% CI, 1.40–2.27; n=4; I2=0.0% and medical management: HR, 1.37; 95% CI, 1.25–1.50; n=4; I2=6.3%; Table II). Patients with LC with RI/CKD had statistically similar DFS, compared to patients with LC and no associated RI/CKD (HR, 1.12; 95% CI, 0.75–1.67; n=3; I2=62.6%; Fig. 3), with no apparent publication bias demonstrated using Egger's test (P=0.204) and the funnel plot (Fig. S2).

Figure 2. Association between renal insufficiency/chronic kidney disease and overall survival. HR, hazard ratio; CI, confidence interval; ES, effect size.

Figure 3. Association between renal insufficiency/chronic kidney disease and disease-free survival. HR, hazard ratio; CI, confidence interval; ES, effect size.

Table II. Overall survival subgroup analysis.

Table II.

Overall survival subgroup analysis.

Variable	n	HR (95% CI)	I2, %
Stage
I/II	5	1.76 (1.30–2.37)a	58.6
III/IV	6	1.18 (0.91–1.54)	64.6
Criteria for calculating eGFR
CKD-EPI	3	1.55 (0.97–2.47)	58.6
ICD-9	3	1.40 (1.14–1.71)a	52.0
Tumor type (non-small cell lung cancer)	7	1.54 (1.22–1.94)a	33.8
Primary management
Surgery	4	1.78 (1.40–2.27)a	0.0
Medical (chemotherapy/radiotherapy/targeted therapy)	4	1.37 (1.25–1.50)a	6.3

a P<0.05. HR, hazard ratio; CI, confidence interval; eGFR, estimated glomerular filtration rate; CKD-EPI, chronic kidney disease-epidemiology collaboration; ICD, International Statistical Classification of Diseases.

Discussion

The results of the present study indicate that RI/CKD is associated with a poor OS of patients with LC, particularly in patients with stage I/II cancer. This finding underscores the importance of identifying renal function status early in the diagnostic process and considering it as a crucial factor in treatment decision-making. The finding related to reduced OS remained consistent regardless of the treatment modality (surgical intervention or medical management). The consistency of the association across different treatment modalities suggests that the influence of RI/CKD on OS is not limited to a specific therapeutic approach. Therefore, renal function should be considered a key factor affecting prognosis and treatment planning regardless of the proposed treatment method. The observed lack of a significant difference in DFS between patients with and without RI/CKD suggests that renal function may have a more pronounced impact on long-term survival rather than on the occurrence of new disease. This result highlights the need for further research into the potential mechanisms behind this association.

Factors such as altered drug metabolism, impaired immune function and increased vulnerability to treatment-related toxicities due to renal dysfunction could contribute to the observed effect on OS (11–13,34,35). Patients with RI or CKD often face challenges in receiving optimal chemotherapy and other treatments due to compromised kidney function (36). Additionally, the usual presence of multiple comorbidities such as cardiovascular disease and diabetes in these patients may exacerbate overall health status and treatment tolerance (37,38). Moreover, at times, delays in cancer diagnosis and treatment initiation may occur due to presence of co-morbidities, including renal dysfunction, potentially allowing the disease to progress to more advanced stages (39,40). Recent studies suggest several potential molecular mechanisms underlying these associations. CKD induces systemic inflammation and immune dysfunction, creating a microenvironment that promotes tumor growth (41,42). Furthermore, impaired renal function can alter the metabolism and pharmacokinetics of anticancer drugs, leading to suboptimal drug levels or increased toxicity, which may affect treatment efficacy (43,44). Additionally, renal dysfunction is associated with increased oxidative stress and DNA damage, processes that contribute to cancer progression and resistance to therapy (45,46).

The observed disparity in findings between studies that used an ICD-9-based diagnosis of CKD and those employing other objective criteria, such as the CKD-EPI, raises questions regarding the influence of diagnostic criteria on the reported associations with survival outcomes. ICD-9 codes are primarily used for administrative and billing purposes and may not capture subtle variations in kidney function that could be relevant for predicting survival outcomes (47). Additionally, ICD-9-based diagnoses may capture more severe cases of CKD, as these codes are often assigned when the condition has progressed significantly enough to necessitate medical attention or intervention. Conversely, the objective criteria of the CKD-EPI equation consider several factors such as age, sex and serum creatinine levels, and may, therefore, be more sensitive and specific in assessing kidney function (31–33). As a result, studies relying on these objective criteria may provide a more accurate representation of kidney function status.

The association observed between RI/CKD and poor OS in studies that specifically focused on NSCLC could have important implications for understanding the relationship between renal function and survival outcomes in this particular cancer subtype. When narrowing the analysis to studies that focused on NSCLC, the sample size may become more homogenous regarding cancer type, treatment modalities and patient demographics. This homogeneity (I2≤40% in the present study) may reduce the confounding effects of these variables, making it easier to detect the influence of RI/CKD on OS. In heterogeneous patient groups that include several cancer types, the effects of renal function may be masked by other factors that differ among the included studies.

The results of the present study may have important clinical implications. RI/CKD is a common comorbidity, especially among the elderly who have associated lung cancer (48). Close monitoring of renal function and potential adjustment of treatment plans to account for altered drug metabolism and potential treatment-related toxicities may help to optimize patient outcomes. However, despite the valuable insights provided by the present meta-analysis, certain limitations should be acknowledged: The included studies were retrospective in design, and the unadjusted confounders may have impacted the results. Heterogeneity across included studies, differences in sample sizes, treatment protocols, tumor grades, histologic types, and variations in the definition of RI/CKD may also have influenced the results. Another limitation is that the medical management of RI/CKD and its potential impact on the association with LC outcomes could not be assessed. Upon reviewing the included studies, it was demonstrated that they primarily focused on reporting LC characteristics and treatments rather than detailed information on the management of RI/CKD itself. Consequently, limited data are available within the reviewed studies to perform a comprehensive analysis of how specific medical treatments for RI/CKD may influence the outcomes of patients with LC. It is of critical importance to investigate the role of RI/CKD management in influencing LC outcomes. Future studies specifically designed to assess the impact of renal treatments, including nephroprotective strategies during cancer therapy and optimization of renal function, would provide valuable insights into improving outcomes for patients with LC with pre-existing renal dysfunction.

In conclusion, the results of the present meta-analysis shed light on the association between RI/CKD and poor OS in patients with LC. The findings underscore the need for a holistic approach to patient care that includes a thorough assessment of renal function early in the diagnostic process and its incorporation into treatment decisions. However, further studies are needed to clarify the mechanisms of this association and to develop approaches that can optimize outcomes for patients with LC with compromised renal function

Supplementary Material

Supporting Data

Supporting Data

Acknowledgements

Not applicable.

Funding

Funding: No funding was received.

Availability of data and materials

The data generated in the present study may be requested from the corresponding author.

Authors' contributions

HQ and SL conceived and designed the study. HQ, SL and ZH were involved in acquisition of data, analysis, and interpretation of data. HQ and SL were involved in the writing of the manuscript and ZH was involved in revising it critically for important intellectual content. All authors have read and approved the final manuscript. HQ, SL and ZH confirm the authenticity of all the raw data.

Ethics approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References