Circulating 25-hydroxyvitamin D Levels and Prognosis among Cancer Patients: A Systematic Review

There is robust evidence that vitamin D is associated with a reduction in cancer risk, particularly breast and colorectal cancers (1–3). Less well known, however, is the impact of vitamin D on prognosis in patients with cancer. Experimental studies have demonstrated that vitamin D can suppress tumor progression and metastasis, via its effect on cellular proliferation, differentiation, and angiogenesis (4)—biologic properties that might be relevant to, and mediate its impact on prognosis in patients with cancer. Vitamin D is converted to 25-hydroxyvitamin D (25-OHD) by cytochrome P450 R21 in the liver (4, 5). Further hydroxylation of 25-OHD to the active form, 1,25-dihydroxyvitamin [1,25-(OH)₂D] by cytochrome P450 27B1 occurs in the kidney (4–7). The 1,25-(OH)₂D generated in the kidney is secreted into the circulation, bound to vitamin D binding protein (VDBP), and then transported to target organs, where it induces genomic and nongenomic responses through its interaction with the vitamin D receptor (VDR; ref. 4). Many cells and tissues express cytochrome P450 27B1 as well as VDR, which implies that local conversion of 25-OHD to 1,25-(OH)₂D, the active form, can take place in such tissues (4).

Nevertheless, experience from the associations of folate with colorectal cancer suggests that bioactive nutrients may have differential effects on tumor initiation and progression, hence, differential effects on cancer incidence and survival (8). Whereas folate deficiency may predispose to colorectal carcinogenesis in normal colorectal epithelial cells, high folate levels seem to promote the growth of an existing cancer (8–10). Hence, understanding the relationship between vitamin and prognosis among patients with cancer is essential, as the impact of vitamin D on cancer risk may differ from its impact on cancer prognosis.

Initial ecological studies indicate a potential inverse association between measures of ultraviolet B (UVB) radiation and cancer mortality (11, 12). Although exposure to UVB is a strong determinant of vitamin D status, it fails to account for differences in vitamin D intake, adiposity, physical activity, and skin pigmentation, which are also important determinants of vitamin D status. Because circulating 25-OHD level is the best indicator of vitamin D status, the associations of vitamin D with prognosis in patients with cancer are best ascertained using circulating 25-OHD concentrations. However, few studies have examined these associations.

An overview of the relationship between 25-OHD concentrations and prognosis in patients with cancer is particularly essential because of the very high number of cancer survivors in the United States (>13 million; ref. 13) and the high prevalence of vitamin D deficiency among these patients (14).

The objectives of this review are to (i) summarize results from studies that have evaluated the associations of circulating 25-OHD concentrations with prognosis among patients with cancer, and (ii) identify gaps and future research needs in this area.

We identified studies published between January 2007 and September 2013 using the following databases: PubMed, OvidSP, BioMed Central, EMBASE, and Scopus. Search terms included the following: “vitamin D,” “25-hydroxyvitamin D,” “calcidiol,” “cancer,” “survival,” “mortality,” “prognosis,” as well as different types of cancer (“breast,” “lung,” “colorectal,” “prostate,” etc.). Search queries in the databases yielded 1,397 articles (Fig. 1). Abstracts were examined for relevance to the topic of study, and full texts were retrieved for applicable abstracts. Reference lists of relevant articles were also examined to identify other studies of interest.

We included studies if they met the following criteria: reported on (i) measurement of circulating 25-OHD, (ii) overall survival, disease-specific survival among patients with cancer or other forms of survival such as distant disease-free survival, disease-free interval, recurrence-free survival, event-free survival, (iii) any other prognosis among patients with cancer, and (iv) published in English. The outcomes were defined as (i) overall survival: time from diagnosis to death as a result of any cause; (ii) cancer-specific survival: time from diagnosis to death from specific cancer; (iii) disease-free survival: time from surgery to the date of the first locoregional recurrence, distant metastasis, detection of a secondary primary tumor, or death from any cause; (iv) disease-free interval: time elapsing between cancer diagnosis and local recurrence and/or lymph node metastasis and/or distant metastasis; (v) distant disease-free survival: time between diagnosis and metastasis at distant sites; (vi) risk of relapse: risk of having local invasive tumor recurrence and/or locoregional lymph node metastasis and/or distant metastasis; (vii) time to progression: time from study entry to disease progression, regardless of the patient's treatment status. Disease progression required 25% increase in measurable tumor or an increase in tumor size in patients whose lesions did not meet criteria for measurable disease; and (vii) time to treatment: the time from diagnosis to disease progression requiring treatment.

Twenty-nine studies met the inclusion criteria. Of these, three articles did not present data on association; hence, data from 26 studies were used in the review. From those 26 articles, we extracted information on study design, age, number of participants and events, types of cancer, effect measures, and confounding factors. Studies were available on the following cancers: breast (N = 7), colorectal (N = 6), lung (N = 4), hematologic (N = 4), prostate (N = 2), skin (N = 2), head-and-neck (N = 2), gastric (N = 1), and pancreatic (N = 1). One study (15) evaluated outcomes across multiple cancer sites (breast, colon, lung, and lymphoma).

Because of the heterogeneity of the studies and the limited data for each cancer type, we performed a systematic review of the effects of circulating 25-OHD levels on prognosis, rather than a meta-analysis.

Between January 2009 and September 2013, six studies investigated the associations of circulating 25-OHD levels with overall survival in patients with breast cancer (refs. 15–20; Table 1). Elevated circulating 25-OHD concentrations were associated with statistically significantly better survival in two studies (15, 19), borderline statistically significant better survival in two (16, 18), and no associations in last two (17, 20). In one of the studies, lower circulating 25-OHD concentrations were associated with reduced survival among women who did not have chemotherapy [HR = 1.15; 95% confidence interval (CI), 1.03–1.27 per 25 nmol/L decrease], but not among those who had chemotherapy (HR, 0.91; 95% CI, 0.75–1.08; P-interaction = 0.06; ref. 18).

Three studies have reported on the associations of circulating 25-OHD with breast cancer survival (15, 17, 19). Circulating 25-OHD was associated with breast cancer survival in two of the other three studies (15, 19), but the effect was limited to postmenopausal (HR, 0.15; 95% CI, 0.03–0.63), and not premenopausal women (HR, 0.93; 95% CI, 0.43–2.02) in one study (19).

Other breast cancer outcomes include distant disease-free survival (16, 18, 19), disease-free survival (19, 21), and relapse (19, 20). Circulating 25-OHD concentrations were associated with longer distant disease-free survival (16, 18, 19). Likewise, 25-OHD was inversely associated with relapse after 3 years (HR, 0.50; 95% CI, 0.29–0.85) and 6 years (HR, 0.25; 95% CI, 0.09–0.70; ref. 19) but not during the first year of treatment (19) or with recurrence (20).

From 2008 till date, five studies have examined the associations of circulating 25-OHD levels with overall survival in patients with colorectal cancer and one study among patients with colon cancer alone (refs. 15, 22–26; Table 2). One of these studies used prediagnostic 25-OHD levels (22), whereas another one used predicted 25-OHD levels (25). Four of the five studies in patients with colorectal cancer reported better overall survival among those with higher 25-OHD levels compared with those with lower 25-OHD levels (22–25), whereas the study in patients with colon cancer did not (15). Elevated prediagnostic 25-OHD levels were associated with better survival among patients with rectal cancer but not among patients with colon cancer (22).

Five studies reported on the associations of circulating 25-OHD with colorectal cancer–specific survival (one on colon cancer–specific survival; refs. 15, 22–25). Elevated 25-OHD concentrations were associated with better colorectal cancer–specific survival in two studies (22, 25).

Circulating 25-OHD concentrations were not associated with time to progression, confirmed response, and disease-free survival in patients with colorectal cancer (23, 26).

Four studies have investigated the associations of 25-OHD with overall survival (refs. 15, 27–29; Table 3). Elevated 25-OHD concentrations were associated with better survival in a Norwegian study (HR, 0.19; 95% CI, 0.12–0.30; ref. 15) and worse survival in a Chinese study (HR, 2.54; 95% CI, 1.01–6.41; ref. 27). Zhou and colleagues reported a beneficial effect of 25-OHD on survival among patients with stages IB and IIB non–small cell lung cancer (NSCLC) but not among those with stage IA tumors (29). Heist and colleagues observed no associations between 25-OHD concentrations and survival in patients with stages III and IV NSCLC as well as among patients with adenocarcinoma and squamous cell carcinoma (28).

Circulating 25-OHD was not associated with recurrence-free survival in the overall analysis but suggestion of an association among patients with stages IB and IIB in a study conducted in Boston, MA (HR, 0.75; 95% CI, 0.45–1.23; P-trend = 0.06; ref. 29).

A few other studies have investigated the associations of circulating 25-OHD levels with prognosis in patients with other cancers, including lymphoma (15, 30), leukemia (31, 32), skin (33, 34), head-and-neck (35, 36), gastric (37), pancreatic (38), and prostate cancers (39, 40) with varying results (Table 4). Elevated 25-OHD levels were associated with better overall survival in patients with gastric cancer and lymphoma. Although elevated 25-OHD was not associated with statistically significant better overall survival among patients with skin cancer, it was associated with better recurrence-free survival (HR, 0.79; 95 %CI, 0.64–0.96; ref. 33). The associations of 25-OHD levels with overall survival in prostate and head-and-neck cancers seem equivocal. Of the two studies each conducted among patients with these cancers, one each reported survival advantage with elevated 25-OHD levels, whereas the other did not. For pancreatic cancer, elevated 25-OHD levels were associated with improved survival among patients with stages III and IV pancreatic cancer but not in the overall analysis.

In this review, we identified 26 studies published between January 2007 and September 2013 that evaluated the associations of circulating 25-OHD levels with prognosis among patients with cancer. Our review suggests that elevated circulating 25-OHD levels may be associated with better overall survival in patients with breast and colorectal cancers. However, the associations of circulating 25-OHD levels with prognosis in patients with other cancer types are less clear due to the few studies that have investigated these associations.

Although the association of circulating 25-OHD levels with cancer risk has been explored in many studies, its association with prognosis among patients with cancer is poorly understood. A previous review of eight studies published in 2011 had provided an initial summary of the role of circulating 25-OHD levels in prognosis among patients with cancer (41). The review included two studies each on breast, colorectal, and lung cancers and one study each on melanoma and prostate cancer. In addition to the eight studies in that review, our review included 18 newly published studies, which is an indication of the strong emerging interest on the role of vitamin D in prognosis among patients with cancer. Because of the increasing number of people living with cancer, the importance of evaluating the associations of vitamin D with outcomes among patients with cancer cannot be overemphasized. Currently, close to 14 million Americans are living with a diagnosis of cancer and the number is expected to increase to 18 million by 2022 (13). At the same time, studies have reported a high prevalence of vitamin D deficiency among patients with cancer. In a recent review of 37 studies, 67% of the patients with cancer had vitamin D insufficiency (25-OHD levels between 25–50 nmol/L; ref. 14). If vitamin D is associated with prognosis among patients with cancer, the potential public health and translational importance could be enormous. Nevertheless, this research area is still in its infancy and more work needs to be done to tease out the effects of vitamin D on prognosis in patients with cancer. Below, we highlight a few areas in this emerging field that deserve more attention if this potential is to be realized.

Only a few studies evaluated the associations of vitamin D with prognosis stratified by tumor characteristics and lifestyle factors; hence, there is paucity of information on factors that modify the associations of 25-OHD with prognosis in patients with cancer. Race, body mass index (BMI), tumor subsite, molecular subtype, as well as stage-stratified analyses are essential as they are important determinants of survival and circulating 25-OHD levels. For instance, non-Caucasian patients with breast cancer have a 2-fold increased risk of vitamin D deficiency compared with Caucasian patients with breast cancer, and obese patients with breast cancer have a 3-fold increased risk of vitamin D deficiency compared with normal-weight patients with breast cancer (42). Circulating 25-OHD levels are lower in patients with localized or regional disease than women with in situ disease (43) and in women with triple-negative breast cancer than women with other breast cancer subtypes (44). The EPIC study demonstrated survival advantage among patients with rectal cancer with elevated 25-OHD levels, but not among patients with colon cancer, which is in line with the null effect reported in a study limited to patients with colon cancer (15). Drake and colleagues demonstrated better survival associated with higher 25-OHD levels among patients with diffuse large B-cell, and T-cell lymphoma but not among other lymphoma subtypes (30). Hence, an in-depth investigation of how tumor characteristics and lifestyle factors modify the associations of vitamin D with prognosis is needed. Only two of the six breast cancer studies reported analyses stratified by stage at diagnosis and hormone receptor status (18, 21), one with BMI (18), and none with race. Of particular importance and interest is to determine whether part of the racial and obesity-related differences in cancer survival can be explained by differences in the 25-OHD levels. It has also been suggested that circulating 25-OHD levels may be a biologic marker for health status among patients with cancer, especially physical activity (45, 46). Patients with cancer with better health status may be more physically active and have a greater exposure to sun, which is an important determinant of circulating 25-OHD levels. Thus, physical activity may confound the association of 25-OHD levels with survival in patients with cancer (45).

Patients with cancer change their lifestyle and eating habits after a diagnosis of cancer, hence, postdiagnostic 25-OHD levels may not be representative of usual levels before cancer diagnosis.

Although nothing is known about individual longitudinal changes in 25-OHD levels after a diagnosis of cancer, the prevalence of vitamin D insufficiency (<50 nmol/L) in patients with cancer can be as high as 67%, compared with 35% within the general population (14, 47, 48). In a study, 74% of patients with breast cancer had vitamin D insufficiency (49), suggesting that prevalence of vitamin D insufficiency is higher among patients with cancer compared with the general population. Furthermore, patients with cancer have been reported to have a 2.5-fold increased odd of having vitamin D insufficiency compared with non-cancer patients within the same geographic area (50).

Likewise, chemotherapy might influence 25-OHD levels (51, 52). Patients with colorectal cancer receiving chemotherapy were three-times more likely to have vitamin D deficiency than those not receiving chemotherapy (52). Similarly, patients with early-stage breast cancer treated with adjuvant chemotherapy had a 32% reduction in 25-OHD levels at day 147 of treatment compared with prechemotherapy levels (51). Although most studies measured 25-OHD levels in pretreatment blood samples, some did not, and in some studies, circulating 25-OHD levels in blood samples collected more than 2 years after diagnosis were related to prognosis. To highlight the importance of timing of blood collection, Vrieling and colleagues reported a statistically significant association between circulating 25-OHD levels and breast cancer survival in analysis conducted using samples taken before initiation of chemotherapy but not in samples taken after initiation of therapy (18). Thus, studies evaluating the associations of circulating 25-OHD with cancer prognosis should consider using pretreatment blood samples, rather than posttreatment samples. In addition, a sizeable number of studies did not take into consideration, or correct for, season of blood draw in their analyses despite the well-documented seasonal differences associated with 25-OHD levels.

There is a wide variability in 25-OHD cutoff points related to prognosis in the studies. Some studies derived their cutoff points based on the distribution within their study population, whereas others used median splits and other categories based on sufficient and insufficient 25-OHD levels. Because of the emerging role of vitamin D in many health outcomes, the definition of vitamin D deficiency has extended beyond its relationship with skeletal health alone. Vitamin D deficiency has been traditionally defined as circulating 25-OHD levels <25 nmol/L because clinical evidence of skeletal diseases such as rickets become manifest below this level (53). Because of emerging evidence of the role of vitamin D in other health outcomes, the concept of vitamin D sufficiency based on 25-OHD levels required to maximize intestinal calcium absorption, prevent secondary hyperparathyroidism, as well as maintain optimal health in at-risk groups has gained a lot of traction over the last decade (4, 54, 55). Circulating 25-OHD levels needed for maximal suppression of parathyroid hormone has been estimated to be between 70 and 80 nmol/L; hence, an expert consensus has adopted 25-OHD levels of 75 nmol/L to indicate vitamin D sufficiency, whereas levels <50 nmol/L indicate insufficiency and intermediate levels indicate relative insufficiency (4, 55–57). Many studies have reported protective effect of 25-OHD on cancer risk at levels indicative of sufficiency (≥75 nmol/L). While some of the studies in our review investigated the associations of 25-OHD with prognosis using sufficient versus insufficient levels, many others did not. Since there is a biologic rationale, future studies on 25-OHD and prognosis among patients with cancer should consider evaluating the associations of vitamin D with prognosis based on sufficient 25-OHD levels. In addition, future studies should consider reporting how a unit increase in circulating 25-OHD (preferably 25 nmol/L) might affect prognosis.

This review suggests that vitamin D may be associated with prognosis in patients with breast and colorectal cancer. Importantly, because low 25-OHD levels among patients with cancer can be corrected, this might pave way for studies evaluating the utility of correcting vitamin D deficiency in the management of patients with cancer. This is especially important given the recent review showing that although circulating 25-OHD levels are associated with improved outcomes in observational studies, results from intervention studies have not confirmed these associations (46). Although vitamin D replacement therapy successfully corrected deficient and suboptimal 25-OHD levels among patients with cancer (58), the impact of this correction on prognosis is unknown. Hence, more observational studies, as well as randomized clinical trials, are needed to further evaluate the associations of vitamin D with prognosis among patients with cancer before vitamin D supplementation among patients with cancer can be considered.

No potential conflicts of interest were disclosed.

A.T. Toriola is supported by the Washington University School of Medicine, Barnes-Jewish Hospital Foundation, and Siteman Cancer Center.