Prognosis in metastatic lung cancer: vitamin D deficiency and depression—a cross-sectional analysis

Abstract

Background

Depression and vitamin D deficiency are common in patients with lung cancer and have prognostic implications in cancer settings. However, their relationship and concomitant survival implications have not been evaluated in patients with metastatic lung cancer specifically. We hypothesised that vitamin D deficiency would be associated with depression and inferior cancer-related survival in patients receiving therapies for stage IV lung cancer.

Methods

This was a cross-sectional analysis of vitamin D, depression and lung cancer characteristics. Vitamin D levels were stratified by level (no deficiency ≥30 units, mild deficiency 20 to 29 units and moderate-to-severe <20 units). Depression was measured by the HospitalAnxiety and Depression Scale-Depression (HADS-D). Survival estimations were made using Cox proportional hazard model and Kaplan-Meier analyses.

Results

Vitamin D deficiency was evident in almost half of the sample (n=98) and was associated with significant depression (HADS-D ≥8) (χ²=4.35, p<0.001) even when controlling for age, sex and inflammation (β=−0.21, p=0.03). Vitamin D deficiency and depression were associated with worse survival and showed evidence of an interaction effect (HR 1.5, p=0.04).

Conclusion

Vitamin D deficiency is associated with depression in patients with metastatic lung cancer. Depression modulates the survival implications of vitamin D deficiency in this population. The role of vitamin D deficiency in cancer-related depression warrants further investigation since both are amenable to treatment. Psychological and nutritional prognostic considerations may help inform treatment paradigms that enhance quality of life and survival.

INTRODUCTION

Depressive symptoms, nutritional deficiencies and inflammation are associated with worsened lung cancer-specific prognoses.^1–3 Vitamin D deficiency has been studied extensively in cancer settings and may be a risk factor for cancer initiation and poor prognosis once cancer is diagnosed.^{4 5} However, the role of vitamin D deficiency in lung cancer specifically is less well studied. Vitamin D and its association with depression and survival in the setting of lung cancer have not been evaluated.

Vitamin D deficiency is common in advanced cancer patients, especially among women and non-White patients, and has been associated with fatigue and anorexia, symptoms which are closely related to depression.⁶ In fact, vitamin D deficiency has been linked to depression in the general population and remains a concern especially for elderly populations where the prevalence of cancer and depression are higher than the general population as well.^{7 8}

This last decade has brought considerable improvements in lung cancer therapeutic options. Despite many recent advances in treatment paradigms, unmet health-related quality of life (HRQOL) needs are still abundant for patients with lung cancer.⁹ Also, social determinates of health undermine cancer care and the gains made with modern therapies even while survival has improved for lung cancer subgroups.¹⁰ Achieving health equity for patients with cancer depends on not only cancer treatment but meeting treatment needs of additional factors that lead to worse survival and quality of life.¹¹ For this reason, the relationship between a general nutritional disturbance such as vitamin D deficiency and depression deserve attention in patients with advancing lung cancer. If there is an underlying relationship in this setting, vitamin D deficiency and depression may be particularly actionable when addressed concomitantly.

Inflammation plays a role in cancer prognostication as well as potentially establishing depressive symptoms in patients with cancer given its underlying relationship.¹² Inflammation has been associated with vitamin D deficiency but the relationship, including directionality and potential underlying causal mechanism, is not clear.¹³ When vitamin D deficiency and inflammation co-exist, this association may be situational and therefore may not be causally related but nonetheless may have prognostic implications. Since a putative relationship between depression and inflammation exists,¹⁴ even in the setting of lung cancer,¹⁵ inflammation is an important covariate or control for any potential underlying association between vitamin D and depression. Understanding the role of inflammation along with vitamin D deficiency and depression may be important for any future interventions.

This study evaluates the association between vitamin D deficiency and depression in patients with metastatic lung cancer undergoing systemic treatments and their relationships with survival while controlling for important covariates. We hypothesised that vitamin D deficiency would be associated with depression and inferior survival outcomes and that survival would be further worsened in lung cancer patients with both vitamin D deficiency and depression.

METHODS AND MATERIALS

The study used a cross-sectional design that was approved by the Memorial Sloan Kettering Cancer Center Institutional Review Board May 2018 under protocol number 18–383 ‘survey of routine markers of inflammation and psychological variables in patients with metastatic lung cancer’. Surveys and routine blood work (eg, C-reactive protein (CRP)) were collected from May 2017 to November 2017 in dedicated thoracic medical oncology clinics as incorporated into standard of care practice and were included in the study analyses. Data for survival were evaluated in November 2019. The data are reported based on the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) recommendations.¹⁶

Participants

Inclusion criteria consisted of patients with a confirmed histologic diagnosis of stage IV lung cancer including non-small cell lung cancer, such as squamous cell carcinoma and non-squamous cell carcinoma (eg, adenocarcinoma and other), and small cell lung cancer who were undergoing active treatment, spoke English and had Eastern Cooperative Oncology Group (ECOG) performance status less than or equal to 2.¹⁷

Patients who were not undergoing treatment for their metastatic lung cancers were excluded so that treatment effects could be compared uniformly across the cohort. Patients with other concomitant cancers were excluded. Patients had to be on active treatment for at least 1 month and had to be more than 1 month from receiving the diagnosis of lung cancer to be included. Findings from this study regarding inflammation and depression and lung cancer type are presented else-where.^{18 19}

Procedure

Participants were approached in oncology clinics and asked to participate in a one-time survey by a treating staff member (eg, nurse practitioner, medical oncologist). Participants filled out the questionnaire containing standardised survey questions either prior to the appointment or during chemotherapy or other treatment infusions. During the same routine oncology appointment, CRP laboratory values were obtained the same day that the questionnaires were completed. Patients were asked to raise any concerns with clinic staff and to notify a staff member if they felt significantly depressed or had suicidal thoughts. Survey results were reviewed with patients during the same visit. Referrals to available psychological services were provided in the survey.

Measures

Medical characteristics and demographics

Medical information gathered from the medical record included disease type, treatment type, line of treatment, length of time with diagnosis (pre and post survey), body mass index (BMI) and antidepressant usage. Demographic information gathered from the medical record included age when survey completed, race/ethnicity, marital status and survival information.

Depression

Depression severity was measured by the Hospital Anxiety and Depression Scale-Depression (HADS-D). The HADS-D has been validated in the lung cancer setting.^{20 21} It is a 14-item symptom rating scale that was developed to identify clinically significant cases of depressive disorders among medically ill patients.²⁰ Physical symptoms are excluded from the HADS-D due to the potential confounding effects of illness on symptoms such as sleep, appetite disturbance and fatigue. Responses are rated 0 to 3 points such that a total score may range from 0 to 21 points. A cut-off of 8 on the HADS-D subscale is most commonly used to identify clinically significant depression, with an average sensitivity and specificity of 0.80.^{20 22}

Vitamin D and inflammation

Routine laboratory data were collected from the date of survey administration and were run in a Clinical Laboratory Improvement Amendments certified lab.²³

Serum 25(OH)D levels were measured by the DiaSorin Liaison immunoassay that detects both 25(OH)D₂ and 25(OH)D₃ and therefore provides the total circulating 25(OH) level. The assay for 25(OH)D has an analytical sensitivity (lower detection limit) of 7 to 5 nmol/L. CRP values were obtained by turbidimetric immuno-assay. Higher CRP levels are consistent with greater inflammation (positive acute phase reactant). Inter-assay and intra-assay coefficient of variation is reliably less than 5%.

Study size

This was a convenience sample that incorporated enough patients to power for an adequate analysis given anticipated effect sizes.²⁴ As such the study incorporated data from 98 patients that would provide 80% power to detect at least a small-to-moderate effect size (ie, above 0.3) with an alpha of 0.05.

Statistical analysis

Statistical procedures were performed using the SPSS V.24 software (SPSS, Chicago, Illinois, 2013) and statistical tests were two-tailed with a 5% significance level. Descriptive analyses were done to assess distribution and central tendency. All categorical data were expressed as n (%). Univariate analyses were conducted to assess relationships with vitamin D levels that included the following: correlation analysis using Pearson or Spearman correlation coefficients depending on normality of distribution; independent t-test for dichotomous predictors and analysis of variance for ordinal or nominal predictors with more than two values. Linear by linear association was used to assess vitamin D levels (under 20, 20 to 29, 30 and above) and meeting depression screening criteria (HADS-D ≥8).Multivariate linear regression analysis was conducted after log transformation of vitamin D levels and CRP. Transformation corrected Vitamin D skewness from 1.786 to 0.029 and Kurtosis from 6.754 to −0.026 and CRP skewness from 3.456 to −0.018 and Kurtosis from 15.209 to −0.767. Survival data (overall survival) were evaluated using Kaplan-Meier method and Cox proportional hazards model, which were controlled for advancing disease and poor nutrition using BMI. Survival curves plotted using the Kaplan-Meier method and analysed using the log-rank sum test. Cox multivariate analysis was performed on potential prognostic factors using stepwise hierarchical linear regression, which included assessment of the interaction term (Vitamin D × depression). The proportional hazards assumption underlying the Cox model was verified by examining the plot of Schoenfeld residuals against time. The HR was expressed as the 95% CI. P values less than 0.05 were considered statistically significant. Very limited amount of data were considered missing and did not affect the primary analyses as such statistical analysis did not account for missing data.

RESULTS

Cohort characteristics

Out of 120 patients, 98 were participated (82% response rate). Patients were 66.3 (SD 9.30) years old on average with 46% displaying vitamin D deficiency (<30 units) (table 1). The average vitamin D level was 32.4 (SD 14.3) with 54% at 30 units or above, 28% between 20 to 29 units and 18% under 20 units. The majority of patients were White (89%), women (63%) with adenocarcinoma (78%) who were receiving their first-line cancer treatments (50%). Overall, the cohort scored 5.2 (SD 3.8) on the HADS-D with 24% meeting clinical criteria for depression (HADS-D ≥8). CRP levels were elevated (≥1 mg/L) in 45% of the cohort with a mean CRP of 2.1 mg/L (SD 3.3), range <0.05 to 22.6 mg/L. The average time with lung cancer was 15 months and patients lived on average for 13 months after survey completion.

Table 1.

Clinical and demographic characteristics of the sample

	Total (n=98)	Association with vitamin D
	M (SD)	r	P value
Age (years) (range 45–86)	66.3 (9.3)	0.14	0.16
Vitamin D level (units) (12–106)	32.4 (14.3)	-	-
C-reactive protein (mg/ml) (<0.05–22.6)	2.1 (3.3)	−0.21	0.04
Depression score (HADS-D) (0–18)	5.2 (3.8)	−0.24	0.02
Time with disease (months) (1–120)	15.4 (17.3)	0.18	0.07
Time alive after survey (months) (1–32)	12.7 (8.0)	0.20	0.06
Vitamin D Level	N (%)	F
Over 30	53 (54%)	62.806	<0.001
20–29	27 (28%)
Under 20	18 (18%)
Gender		t	P value
Male	35 (%)	0.506	0.61
Female	63 (%)
Race/ethnicity
Non-White	11 (11%)	0.006	0.99
White	87 (89%)
Married
Yes	68 (70.7%)	0.487	0.63
No	30 (29.3%)
Antidepressant
Yes	19 (19%)	−0.848	0.41
No	79 (81%)
Disease type		F	P value
Adenocarcinoma	76 (78%)	0.937	0.58
Squamous cell carcinoma	4 (4%)
Small cell lung cancer	14 (14%)
Unspecified	4 (4%)
Treatment type
Chemotherapy	34 (35%)	0.947	0.56
Immunotherapy	30 (31%)
Targeted therapy	21 (21%)
Missing	2 (2%)
Line of treatment
First	50 (%)	1.050	0.43
Second	30 (%)
Third or beyond	10 (%)

HADS-D, Hospital Anxiety and Depression Scale-Depression.

Association with depression

Vitamin D levels were inversely correlated with depression (r_s=−0.24, p=0.02) and inflammation (r_s=−0.21, p=0.04) (table 1). Vitamin D categories (<20, 20 and 29, 30 and above) were associated with meeting depression screening criteria (HADS-D ≥8) (linear by linear association, 4.350 p=0.03) (table 2). As expected, depression and inflammation were correlated (r_s=0.29, p=0.001). A partial correlation was run to determine the relationship between vitamin D level and depression while controlling for inflammation. This resulted in a negative partial correlation between vitamin D level and depression that was statistically significant, r(95) = −0.203,p=0.04.The zero-order correlation showed a similar statistically significant, negative correlation coefficient between vitamin D levels and depression (r_s=−0.24, p=0.02) indicating that inflammation (CRP) had relatively little influence in controlling for the relationship between vitamin D and depression. Furthermore, a multivariate model revealed that vitamin D deficiency (β=−0.21, p=0.04) and CRP (β=0.23, p=0.02) independently predicted for depression while controlling for age and sex in this cohort of patients with metastatic lung cancer (table 3).

Table 2.

Relationship between vitamin D levels (<20, 20 to 29 and ≥30) and clinically significant depression

	Depression criteria (HADS-D ≥8)
Vitamin D level	No	Yes	Total
≥30	47	6	53
20 to 29	22	5	27
<20	12	6	18
Total	81	17	98

HADS-D, Hospital Anxiety and Depression Scale-Depression.

Table 3.

Vitamin D and covariate associations with psychological symptoms

Variable	Depression (HADS-D)
	Regression coefficient	Standardised coefficient	t value	P value	95% CI
Control					Low	High
Age	0.01 (0.04)	0.02	0.164	0.87	−0.067	0.079
Sex	−0.12 (0.74)	−0.02	−0.158	0.88	−1.576	1.344
Covariate
CRP (log)	1.30 (0.57)	0.23	2.286	0.02	0.170	2.419
IV
Vitamin D (log)	−4.13 (2.0)	−0.21	−2.026	0.04	−8.170	−0.083
	F 3.037, adjusted R2 0.077, p=0.02

CRP, C-reactive protein; HADS-D, Hospital Anxiety and Depression Scale-Depression.

Survival analysis

Cox regression modelling revealed that survival was almost three times improved with increasing vitamin D levels (HR .14, p=0.01) or three times worsened in patients with vitamin D deficiency while controlling for advancing disease by incorporating BMI (table 4). When depression was added, the model did not change significantly (χ² change 1.018, p=0.31) but there was a significant interaction effect (depression × vitamin D) (HR 1.5, p=0.04) changing the model significantly (χ² change 4.284, p=0.04).

Table 4.

Cox proportional hazards hierarchical regression model evaluating the influence of vitamin D levels and depression on survival

Model 0	Survival			CI
	Regression coefficient	HR	P value	Lower	Upper
Age	0.00 (0.01)	1.002	0.87	0.975	1.031
Sex	−0.16 (0.29)	0.855	0.59	0.481	1.521
BMI	−0.05 (0.03)	0.949	0.09	0.894	1.008
	χ2 change 3.121, p=0.37
Model 1	Survival			CI
	Regression coefficient	HR	P value	Lower	Upper
Age	0.01 (0.01)	1.009	0.54	0.981	1.037
Sex	−0.21 (0.30)	0.811	0.48	0.453	1.452
BMI	−0.06 (0.03)	0.942	0.05	0.886	1.001
Vitamin D (log)	−1.96 (0.76)	0.141	0.01	0.030	0.655
	χ2 change 6.228, p=0.01
Model 2	Survival			CI
	Regression coefficient	HR	P value	Lower	Upper
Age	0.01 (0.01)	1.007	0.64	0.979	1.036
Sex	−0.22 (0.30)	0.807	0.47	0.45	1.444
BMI	−0.06 (0.03)	0.939	0.04	0.883	0.998
Vitamin D (log)	−1.77 (0.82)	0.171	0.03	0.034	0.851
Depression	0.03 (0.03)	1.029	0.30	0.975	1.085
	χ2 change 1.018, p=0.31
Model 3	Survival			CI
	Regression coefficient	HR	P value	Lower	Upper
Age	0.01 (0.01)	1.006	0.69	0.978	1.034
Sex	−0.22 (0.29)	0.803	0.46	0.451	1.430
BMI	−0.06 (0.03)	0.941	0.04	0.887	0.998
Vitamin D (log)	−3.99 (1.41)	0.018	0.005	0.001	0.294
Depression	−0.56 (0.30)	0.570	0.06	0.318	1.025
Vitamin D × depression	0.42 (0.21)	1.522	0.04	1.009	2.296

χ² change 4.284, p=0.04.

BMI, body mass index.

Patients with vitamin D deficiency (<30 units) had a median survival estimate of 312 days (SE 92) (95% CI 130 to 494 days). Patients without vitamin D deficiency had a median survival estimate of 611 days. The overall group estimate for mean survival was 496 days (SE 32.9) (95% CI 431 to 560 days) (log-rank χ² 4.086 (p=0.04)) (figure 1). In addition, higher vitamin D levels trended towards correlation with survival by two parameters: time with disease (at the time of survey) (r_s=0.18, p=0.07) and time alive after survey (r_s=0.20, p=0.06).

Figure 1.

Kaplan-Meier plots of patients with vitamin D deficiency (<30 units) versus no vitamin D deficiency (≥30 units).

When meeting depression criteria was factored into vitamin D deficiency the following group differences for survival emerged:

• Group 0 (no vitamin D deficiency + no depression): 557 days (SE 45) (95% CI 468 to 645 days); median survival 611 days; n=47, 22 events, 25 censored.

• Group 1 (depression but no vitamin D deficiency): 461 days (SE 122) (95% CI 122 to 221 days); median survival 240 days; n=6, 3 events, 3 censored.

• Group 2 (vitamin D deficiency but no depression): 453 days (SE 55) (95% CI 346 to 561 days); median survival 340 days; n=34 with 21 events and 13 censored.

• Group 3 (depression + vitamin D deficiency): 355 days (SE 99) (95% CI 160 to 550 days); median survival 192 day; n=11 with 7 events and 4 censored.

Pair-wise analysis revealed that differences in survival were only significant for groups 0 and 3 (log-rank χ² 4.394 (p=0.04)) while the rest of the relationships were not significant (figure 2).

Figure 2.

Kaplan-Meier curve of patients with or without vitamin D deficiency (<30 units) or depression (Hospital Anxiety and Depression Scale ≥8). (0) No vitamin D deficiency and no depression; (1) depression but no vitamin D deficiency; (2) vitamin D deficiency but no depression; and (3) both vitamin D deficiency and depression.

DISCUSSION

While this study is exploratory in nature, the results are hypothesis generating and suggest that vitamin D deficiency and depression are not only associated in this sample of patients with lung cancer but their combined presence may be multiplicative in terms of overall survival. The association between vitamin D and depression was similar in magnitude to the association between inflammation and depression, a known and continuously reported association. The multivariate model demonstrated a relatively similar contribution of vitamin D deficiency and inflammation in the prediction of depressive symptoms. In other words, vitamin D deficiency independently predicts for depression regardless of inflammation. This is significant because inflammation is a known predictor of depression in the cancer setting and vitamin D deficiency appears to have similar importance and may have treatment implications.

This study sheds light on vitamin D deficiency in a relatively selective cohort of cancer patients and evaluates its associations with another known adverse factor, depression and the important outcome of survival. While previous studies have evaluated the association between vitamin D deficiency and depression in non-medically ill, healthy populations, relatively few have evaluated the relationship in the context of a specific cancer type. This study suggests that there is a relationship between vitamin D deficiency and depression in patients with lung cancer, which is reflected by their interaction leading to worse survival in the presence of concomitant depression and vitamin D deficiency. Not only is lung cancer associated with high levels of vitamin D deficiency and depression but they are related and may contribute to worsened survival outcomes despite treatment advances.

Studies of vitamin D supplementation have shown responses that may be selective for particular subgroups, for example, patients with early stage lung cancer.²⁵ At the same time, micronutrient deficiencies (eg, B12, folic acid, vitamin D deficiencies) are common in psychiatric disorders and may contribute to refractory conditions and suicidal behaviours.^{26 27} A meta-analysis revealed a positive effect of vitamin D supplementation on depression measures in other-wise health patients.²⁸ However, studies have not yet emerged for patients with cancer who have both vitamin D deficiency and depression. Treatment of cancer-related depression may benefit substantially from adequate vitamin D supplementation.

The implications of this study may be particularly relevant for elderly populations who are frequently living with both vitamin D deficiency and cancer and suffer from high rates of depression as well. The role of vitamin D deficiency and depression in cancer patients is not defined, especially in those elderly patients with lung cancer. Metabolic disturbance is a hallmark of cancer and therefore hyper-metabolism may be a pathway to vitamin D deficiency.²⁹ At the same time, many patients who develop lung cancer may have had premorbid vitamin D deficiency prior to the cancer diagnosis. Regardless of the mechanism, the potential association with depression is concerning given the HRQOL and survival implications. Both depression and vitamin D deficiency are reversible and could be addressed concomitantly. Also, vitamin deficiencies and neurometabolic disorders have been linked to suicidal behaviour, which is unfortunately more prevalent in patients with cancer and particularly high in patients with lung cancer.²⁷ In fact, the relative risk of suicide is elevated to 13 times in the initial weeks to 3 months after a new cancer diagnosis, which continues to be 3 to 5 times risk for the first year of diagnosis.^30–32 Therefore, addressing vitamin D deficiency at the diagnosis of lung cancer may ameliorate depressive symptoms during the lung cancer trajectory. These findings warrant further investigation.

Limitations

These findings are correlational in nature with no purported directionality. In other words, there may be a hidden variable that explains how vitamin D deficiency and depression are related even though age, sex and BMI (surrogate for advancing disease) were controlled and there were no differences in vitamin D deficiency based on lung cancer or treatment type. However, it is possible that this association simply identifies sicker patients with metabolic derangements who will die sooner, for example. However, the other control for disease progression was inflammation. There may be another variable that links these factors together despite controlling for advancing disease with inflammation and BMI. This study was cross-sectional at a single institution that is a cancer referral centre and not necessarily representative of community oncology practices. Also, the study did not collect information on reasons for non-participation. Therefore, the results should be considered exploratory but hypothesis generating.

Summary

Patients with metastatic lung cancer have high levels of both vitamin D deficiency and depression. The combination highlights at-risk patients who may require a higher level of supportive care. Depression interventions in the cancer setting may benefit from adjuvant vitamin D supplementation, which has the potential benefit of also improving cancer-related survival.