• folate;
  • head and neck cancer;
  • laryngeal cancer;
  • molecular epidemiology


  1. Top of page
  2. Abstract


It has been previously observed that patients with head and neck squamous cell carcinoma or with laryngeal leucoplakia present a significant reduction in plasma folate levels. The current Phase IIA pilot study assessed the effectiveness of folic acid as a chemopreventive agent in patients affected by glottic laryngeal leucoplakia,


Forty-three untreated patients affected by glottic laryngeal leucoplakia were enrolled in the Ear, Nose, and Throat Department (Universita Cattolica del Saco Cuore, Rome, Italy). Glottic leucoplakia was initially diagnosed by indirect laryngoscopy and successively confirmed by diagnostic direct microlaryngoscopy with a biopsy for histologic assessment. Folic acid (Folina, Schwarz Pharma, Germany) was administered orally (5 mg every 8 hours) for 6 months. Patients were monitored every 30 days by videolaryngoscopy.


Twelve (28%) patients had no response, 19 (44%) had a partial response, and 12 (28%) had a complete response. The mean increase in serum folate levels (10.06 ± 0.53) and the mean decrease in homocysteine serum (3.65349 ± 0.85526) at the end of the study were highly significant (P = .0001).


The larynx is 1 of the sites of major interest and a good model for the development of chemopreventive agents, but so far the proposed agents have shown no clear efficacy on precancerous lesions or on the development of second malignancies. Cancer 2006. © 2006 American Cancer Society.

Head and neck cancer squamous cell carcinomas (HNSCCs) include SCCs of the oral cavity, pharynx, and larynx. Approximately 39,000 cases of HNSCC were estimated to occur in the US in 2004, with 11,000 deaths.1 HNSCCs represent about 3% of all malignant tumors in the US.1 However, in other parts of the world, such as India, Southeast Asia, and Brazil, the disease is much more prevalent.2 The standard therapeutic approach, focusing on surgery, irradiation, and chemotherapy, alone or in combination, has been modified in part over the last 30 years, but the overall survival of HNSCC patients has not improved substantially.1, 2 Efforts toward early detection and prevention have not been entirely successful. For patients affected by early-stage carcinomas, with a high disease-specific survival rate, second primary tumors represent the first cause of death.2–5

The best-established risk factors for HNSCC are behavioral ones, and primary prevention can be easily obtained by suspension of harmful habits. Smoking and alcohol intake are the best-defined risk factors for HNSCC. A role has also been proposed for low vegetable intake,6 infections by Epstein-Barr virus (especially for nasopharynx) and human papillomaviruses (for oropharynx and other sites),7, 8 betel quid chewing,9 marijuana smoking,10, 11 and acid and alkaline reflux.12, 13 However, it is known that not all laryngeal squamous cell carcinoma (LSCC) patients have a history of behavioral risk factors. To characterize and thus identify high-risk mucosal areas and preclinical tumors, molecular abnormalities in head and neck carcinogenesis have been studied extensively.14–16 As a genetic predisposition to the development of HNSCC is highly probable, in the last years a number of genetic polymorphisms have been evaluated in relation to the risk of developing cancers of the aerodigestive tract.17–19 Metabolic aspects in head and carcinogenesis have been less extensively studied. Nevertheless, we know that metabolic alterations, often aspecific, are frequently associated with cancer. These may be secondary or may precede tumor development and favor progression.

Folate, a water-soluble B vitamin, which is plentiful in fresh vegetables and fruits, is present under a number of coenzymatic forms, whose main biochemical function in mammalian systems is to mediate the transfer of 1-carbon units at different states of oxidation. It is fundamental in the synthesis of serine from glycine, in the synthesis of purine and pyrimidine bases, and as a methyl donor to create methylcobalamin, which is used for remethylation of homocysteine to methionine (methionine cycle).20 Folate deficiency is almost always secondary to an insufficient dietary intake.

Maintenance of adequate folate status from dietary sources and/or by synthetic folic acid supplementation has been associated with a protective effect and reduced incidence of a variety of human cancers.21 In a large perspective study, a high dietary folate intake was shown to protect, in particular heavy smokers, from lung squamous cell carcinoma development.22 Furthermore, folate and vitamin B12 supplementation has been described to induce the regression of bronchial squamous metaplasia.23, 24 An inverse correlation between folate intake and pancreatic and breast cancer risk was also reported.25, 26 A recent animal study evidenced a strong protective effect of folate in beagles treated with a gastric carcinogen.27 A role for folate deficiency, first postulated in the 1960s,28 and for other alterations of methionine cycle metabolites (homocysteine, vitamin B12) as a risk factor for cervical cancer was never definitely demonstrated.29–32 Colorectal carcinogenesis was extensively studied in relation to folate metabolism and several epidemiologic studies showed an increased risk associated with folate deficiency. Patients affected by colorectal cancer seem to have a lower folate dietary intake,33, 34 higher homocysteine, and lower folate serum levels.35 A role of folate deficiency in colorectal carcinogenesis is confirmed also by animal studies.36, 37 Folate is at present considered 1 of the most promising chemopreventive agents for colorectal carcinogenesis.38

Several cellular effects have been described that may account for the potential role of hypofolatemia in carcinogenesis. Methionine cycle disruptions, by reducing intracellular S-adenosylmethionine (SAM), can alter cytosine methylation in DNA, leading to inappropriate activation of protooncogenes, repression of tumor suppressor genes, and induction of malignant transformation. Alterations in DNA methylation and in particular a global hypomethylation and a regional hypermethylation, especially of promoters of tumor suppressor genes, have been described in human tumors.39 In head and neck cancer, promoter hypermethylation of key genes in critical pathways, as INK4A, is common and has also been recently described.40 Alternatively, abnormal DNA metabolism and a variety of cytogenetic lesions have been associated with folate deficiency in laboratory models as well as in human folate deficiency.41, 42 Normal levels of the precursor nucleotides (dNTPs and NTPs) for DNA/RNA synthesis are directly dependent on intracellular folate availability43 and dNTP pool imbalance seems of great relevance for the carcinogenic effects of folate deficiency.44 A specific and important biochemical alteration, secondary to a defect in dTMP synthesis, is uracil misincorporation in DNA, which might be sufficient to determine double-strand breaks, mutations, and chromosomal aberrations.45, 46 On the whole, dNTP pool imbalance reduces the efficiency of both DNA-synthesizing enzymes, with an increase in the background mutation rate, and of DNA-repairing enzymes, with an enhancement of mutagen-induced carcinogenesis.21, 47 Convincing evidence from clinical studies indicates that moderate folate deficiency would not be independently mutagenic in vivo, but probably interacts with other risk factors in promoting tumor progression. In fact, studies about lung and cervical carcinogenesis suggest that folate deficiency enhances an underlying predisposition due to environmental factors, such as heavy cigarette smoking and human papillomavirus (HPV) infection.48, 49 We previously observed that patients with HNSCC or with laryngeal leucoplakia present a significant reduction in plasma folate levels compared with controls.50, 51 Thus, monitoring of plasma folate levels can be used to define a subset of high-risk subjects; furthermore, folate supplementation, whether obtained via the dietary or the pharmacologic approach, could be considered a novel tool for chemoprevention. Indeed, an increase in recommended dietary allowances (RDA) for folate was already proposed,52 and a strong protective effect of folate against gastric carcinogenesis has been recently reported in an experimental animal model.27 It is also well established that pharmacologic folate supplementation, even at fairly high dosage, as occurs in the case of malabsorption, is virtually devoid of any known side effect. The above considerations provide a rationale for clinical investigation in humans. Chemoprevention was defined as an attempt to reverse, suppress, or delay the progression from normal mucosa toward invasive cancer.53 In the present work we performed a Phase IIA pilot study to assess the effectiveness of folic acid as a chemopreventive agent in patients affected by glottic laryngeal leucoplakia, a relatively frequent precancerous lesion.


  1. Top of page
  2. Abstract


Forty-three untreated patients affected by glottic laryngeal leucoplakia were enrolled in our Ear, Nose, and Throat Department (Universita Cattolica del Sacro Cuore, Rome, Italy) in the years 2002-2003. Table 1 summarizes information collected regarding patients' behavioral risk factors.

Table 1. Patient Characteristics
 Laryngeal leukoplakias
Mean age (range), y54 (28-77)
Gender (F/M)3/40
Smokers (%)38 (88.4)
 10 to 20 cigarettes daily10 (23.3)
 >20 cigarettes daily28 (65.1)
 Smoking period <20 years19 (44.2)
 Smoking period >20 years19 (44.2)
Drinkers (%)30 (69.8)
 ≤2 glasses of wine/day20 (46.5)
 2 to 4 glasses of wine/day10 (23.3)

An exclusion criterion from our study was estimated habitual alcohol consumption higher than 35 g of alcohol per day. In fact, it has been reported that, whereas a low to moderate alcohol consumption does not appear to determine any change in serum levels of folate and homocysteine, heavy alcohol drinking can alter folate absorption and metabolism.54 No subject included in the study had received folate or vitamin B12 supplements in the last 6 months. All patients had normal renal function. None of the patients included in the study had a diagnosis of laryngopharyngeal reflux disease (LPRD), or was undergoing treatment with drugs effective in LPRD as proton pump inhibitors, which could have been a confounding factor as there is a great deal of evidence suggesting that this clinical problem can contribute to the development of laryngeal leukoplakia and dysplasia as well.12, 13

Other exclusion criteria were 1) a previous diagnosis of malignancy (because of the hypothetical risk that folate increased DNA synthesis and the proliferation rate of neoplastic cells); and 2) patients who agreed to quit smoking, which is currently considered as the only preventive measure demonstrated to be effective and might have been a bias for our study.

Study Design

The following study protocol was approved by the Ethical Committee of Catholic University Medical School in Rome. Glottic leucoplakia was initially diagnosed by indirect laryngoscopy. The majority of spontaneous clinical regressions of glottic leucoplakia occur in the first month after diagnosis; therefore, in the present study we enrolled only patients with persistent leucoplakia, who normally undergo surgery. We confirmed the diagnosis by diagnostic direct microlaryngoscopy, with a biopsy for histologic assessment. We performed the biopsy by taking a small sample with microcup forceps. In fact, glottic leucoplakia can be associated with different histologic patterns. Leucoplakia was histologically classified as normal mucosa, hyperkeratosis without dysplasia, mild dysplasia (LIN I), moderate dysplasia (LIN II), severe dysplasia (LIN III), or invasive carcinoma, as previously reported.55–57 All leucoplakias diagnosed as invasive carcinomas were excluded from the trial, and patients underwent laser excision or were irradiated with curative intent. The remaining patients were enrolled in our study; all provided an exhaustively informed consent.

A baseline histologic score from 1 (normal mucosa) to 5 (severe dysplasia) was assigned. Thereafter, folic acid (Folina, Schwarz Pharma, Germany) was administered orally (5 mg every 8 hours) for 6 months. Before starting the treatment, we performed and recorded a videolaryngoscopy and assessed serum folate and homocysteine levels. Patients were monitored every 30 days by videolaryngoscopy (armed follow-up). Folatemia and homocysteinemia were also assessed monthly in order to verify the metabolic response and to obtain information about patient compliance. Response to the treatment was always assessed based on the videolaryngoscopy findings by 3 different experienced laryngologists, because leucoplakia is a clinical entity. In case of disagreement, consensus was reached by joint reevaluation of the videolaryngoscopy records. The response was then classified as null (response score = 1) when the dimensions and appearance of the lesion remained substantially unmodified, as partial (response score = 2) when a dimensional reduction at videolaryngoscopy of 50% or more was observed, or as complete (response score = 3) when there was no evidence of residual leucoplakia. If during treatment the lesion had subjectively worsened, or the dimensions increased, we would have resected it; in the present study this never occurred. If the lesion persisted after 6 months, the patients underwent endoscopic resection followed by a second histologic assessment and scoring of the residual leucoplakia.


Homocysteine was measured by a fully automated AxSYM method (Abbott Diagnostics, Abbott Park, IL) according to the manufacturer's directions. The automatic method is based on the determination of S-adenosyl-L-homocysteine (SAH), obtained from the enzymatic conversion of reduced homocysteine to SAH by bovine SAH hydrolase. The Abbott AxSYM immunoassay is based on fluorescence polarization immunoassay (FPIA) technology. After the addition of mouse monoclonal SAH antibody to the sample, S-adenosyl-L-cysteine fluorescein tracer, which competes with SAH for antibody binding sites, was added. The homocysteine concentration was then quantified by the intensity of polarized fluorescent light. Serum folate was measured by ion capture assay on an AxSYM Analyzer (Abbott Diagnostics).

Statistical Analysis

We performed statistical analysis with JMP software (v. 5.1; SAS Institute, Cary, NC).


  1. Top of page
  2. Abstract

The histologic diagnosis at baseline was: normal mucosa in 5 (11.6%) cases, hyperplasia in 11 (25.6%) cases, mild dysplasia in 9 (20.9%) cases, moderate dysplasia in 9 (20.9%) cases, and severe dysplasia in 9 (20.9%) cases. The lesion was monolateral in 31 (72%) and bilateral in 12 (28%) patients.

With regard to smoking and drinking habits, none of the patients reported any change during the study or in the previous year.

Twelve (28%) patients had no response, 19 (44%) had a partial response, and 12 (28%) had a complete response (Fig. 1).

thumbnail image

Figure 1. The complete response of patients M.G. (A) before treatment and (B) after treatment; and Q.C. (C) before treatment and (D) after treatment.

Download figure to PowerPoint

If we exclude from our analysis the 5 patients with a normal histology of the biopsy, only 7 patients (18.5%) had no response, and the complete response and partial response rates rise to, respectively, 31.5% and 50%.

When a complete response was not obtained (n = 31), in most cases (n = 20) the histology of residual leucoplakia was unchanged from the beginning of the study; in 8 cases (Fig. 2) we observed a histologic regression (with a lower final histologic score); in 3 cases we observed a histologic progression, but in these patients the final score was never higher than 2, thus indicating a (presumably innocuous) passage from normal mucosa to hyperplasia. On the whole, the histologic diagnosis of residual lesions was: normal mucosa in 2 (6.5%) cases, hyperplasia in 16 (51.6%) cases, mild dysplasia in 9 (29%) cases, moderate dysplasia in 3 (9.7%) cases, and severe dysplasia in 1 (3.2%) case (Table 2). The end response and the histologic trend are shown in relation to the histology at diagnosis in Figure 3. In particular, the percentage of clinical partial and complete responders rises, respectively, to 44% and 50% if we consider only patients with advanced laryngeal premalignancy (moderate to severe dysplasia). As for histology, among not complete responders we observed 66% of patients with histologic regression, 33% with unvaried histology, and no cases of histologic progression during the 6 months of folic acid supplementation (Table 2; Fig. 3).

thumbnail image

Figure 2. Histologic regression in a case of partial response to folate supplementation (A) before treatment and (B) after treatment.

Download figure to PowerPoint

Table 2. Patients Sorted by Histology, End Response, and Histologic Trend (With Initial Folate Levels, Initial Histologic Score, and End Histologic Score)
 HistologyEnd ResponseInitial Folate Level, ng/mlHistologic Score*End Histologic ScoreHistologic Trend
  • *

    Histologic scores are defined in Study Design.

1NormalNo response12.612Histologic progression
2NormalNo response8.512Histologic progression
3NormalNo response5.312Histologic progression
4NormalNo response7.611Same histology
5NormalNo response5.211Same histology
6HyperplasiaComplete resp5.520Complete regression
7HyperplasiaNo response422Same histology
8HyperplasiaNo response5.722Same histology
9HyperplasiaNo response6.222Same histology
10HyperplasiaNo response6.822Same histology
11HyperplasiaNo response722Same histology
12HyperplasiaNo response4.622Same histology
13HyperplasiaPartial response6.322Same histology
14HyperplasiaPartial response8.422Same histology
15HyperplasiaPartial response5.122Same histology
16HyperplasiaPartial response6.822Same histology
17Mild dysplasiaComplete response4.130Complete regression
18Mild dysplasiaComplete response4.630Complete regression
19Mild dysplasiaPartial response4.933Same histology
20Mild dysplasiaPartial response5.233Same histology
21Mild dysplasiaPartial response5.333Same histology
22Mild dysplasiaPartial response3.233Same histology
23Mild dysplasiaPartial response533Same histology
24Mild dysplasiaPartial response2.732Histologic regression
25Mild dysplasiaPartial response3.732Histologic regression
26Moderate dysplasiaComplete response7.640Complete regression
27Moderate dysplasiaComplete response340Complete regression
28Moderate dysplasiaComplete response4.240Complete regression
29Moderate dysplasiaNo response3.644Same histology
30Moderate dysplasiaPartial response5.744Same histology
31Moderate dysplasiaPartial response4.342Histologic regression
32Moderate dysplasiaPartial response4.443Histologic regression
33Moderate dysplasiaPartial response3.143Histologic regression
34Moderate dysplasiaPartial response643Histologic regression
35Severe dysplasiaComplete response3.950Complete regression
36Severe dysplasiaComplete response4.950Complete regression
37Severe dysplasiaComplete response450Complete regression
38Severe dysplasiaComplete response4.250Complete regression
39Severe dysplasiaComplete response4.350Complete regression
40Severe dysplasiaComplete response4.550Complete regression
41Severe dysplasiaPartial response3.955Same histology
42Severe dysplasiaPartial response353Histologic regression
43Severe dysplasiaPartial response4.354Histologic regression
thumbnail image

Figure 3. (A) The end response and (B) the histologic trend are shown in relation to the histology at diagnosis.

Download figure to PowerPoint

The median follow-up after the end of the protocol was 12 months (range, 8-18 months). In the follow-up period, 3 patients had a recurrence of laryngeal leucoplakia, 1 had a complete response but developed another lesion after 15 months. The other 2 patients recurred after surgical excision of the residual lesion (previous response score = 1). The histology of the recurrent lesions was always moderate dysplasia, both in the 2 previously operated patients and in the patient with a complete response after folic acid supplementation.

Similar to our previous results in patients with laryngeal leucoplakia,51 the mean serum folate level was 5.1906977 ng/mL (±1.8410419) and the mean serum homocysteine level was 10.902326 μmol/L (±6.2308942). As previously reported,58 homocysteinemia inversely correlates with folatemia (Spearman Rho = −0.57, P<.0001).

The mean increase in serum folate levels at the end of the study was 10.06 ± 0.53, with a high statistical significance on the t-test for matched pairs (P<.0001). In all of the patients we observed an increase in folate levels after treatment. Therefore, the overall compliance of patients was presumably good. The mean decrease in homocysteine serum levels was not as evident (3.65349 ± 0.85526) as reported in literature after folate supplementation,58 but was nevertheless highly significant (P = .0001 at t-test for matched pairs).

A post-hoc analysis on the age of patients was carried out establishing a cutoff at 45 years: an “elderly group” (n = 35) and a “young group” (n = 8) were defined. The Wilcoxon/Kruskal–Wallis test showed that elderly patients present a significantly higher response score (Z = −2.06; P = .0397).

We also performed a Spearman Rho multivariate nonparametric analysis on all the numeric variables (age, folate homocysteine, and B12 serum levels at the beginning and end of the study, histologic score at the beginning and end of the study, response score, sides involved). The following relations emerged:

  • Serum folate levels at the beginning of the study are inversely related to age (Spearman Rho = −0.3; P = .049).

  • The histologic score positively correlates with age (Spearman Rho = 0.3744; P = .0134) and baseline homocysteine levels (Spearman Rho = 0.3392; P = .0261) and inversely correlates with baseline folate levels (Spearman Rho = −0.6155; P<.0001).

  • The response score is positively related to grading (Spearman Rho = 0.6829; P<.0001) and to initial homocysteine levels (Spearman Rho = 0.3262; P = .0328) (Table 2; Fig. 3) and inversely related to initial folate levels (Spearman Rho = −0.3727; P = .0138). Thus, patients with hypofolatemia at entrance in the study were more responsive to folic acid supplementation.


  1. Top of page
  2. Abstract

Among the most frequent malignancies in the US, only laryngeal cancer and cancer of the uterine corpus showed no increase in 5-year survival rates during the last 30 years.1 Several potential reasons for this failure can be identified,51 including the high incidence of second primary tumors (SPTs). If one excludes locoregional disseminations, SPTs, which occur at an annual rate of 4% to 7%, are the first cause of death in patients with early stage LSCCs4, 59, 60 and in particular with early glottic cancers,61 characterized by a high rate of locoregional control.

Selected high-risk populations, such as heavy smokers and drinkers, present a high incidence of premalignant lesions (i.e., leucoplakia, erythroplakia) in laryngeal mucosa, most frequently in the glottic region (true vocal folds). In particular, laryngeal leucoplakia can be associated with heterogeneous histologic patterns: normal mucosa, hyperplasia, dysplasia (from mild to severe), and even invasive carcinoma. The reported frequency of progression from dysplastic leucoplakia to invasive carcinoma varies among the studies, but it is always related to the histology of the lesion, ranging from 3% to 5% for hyperplasia and mild dysplasia, to 205 to 30% for severe dysplasia.55 This pattern is consistent with the current concept of a temporal and spatial multistep progression toward malignancy within the “field of cancerization,”62, 63 which progresses with age64 and even decades before a malignant phenotype emerges.14 A standard treatment for leucoplakia and other dysplastic lesions has not been established, and the proposed approaches range from observation to surgical resection.65 Currently, none of these therapeutic options appears to effectively prevent recurrence or even transformation.55

The above considerations make the larynx 1 of the sites of major interest and a good model for the development of chemopreventive agents. Furthermore, the laryngeal mucosa is easily accessible by laryngoscopy, and it is therefore relatively simple to diagnose premalignant lesions and to assess the response to a chemopreventive therapy.

The agents proposed so far show no clear efficacy on precancerous lesions, or on the development of second malignancies. In spite of a noninnocuous toxicity profile, retinoids have been used as chemopreventive agents, and encouraging results have been reported in the treatment of laryngeal66 and oral67 precancerous lesions. However, the large-scale trial EUROSCAN resulted in no benefit—in terms of survival, event-free survival, or second primary tumors—to patients with HNSCC or lung cancer after a 2-year supplementation with retinyl palmitate and/or N-acetylcysteine.68

The present work was aimed at a preliminary evaluation of folic acid as a novel chemopreventive agent. Definitive conclusions are far from being drawn because reliable data about spontaneous regression of laryngeal leucoplakia are lacking. Our complete response rate is lower than the 1 reported in a smaller population (23 patients) after treatment for 12 months with a 3-drug regimen of isotretinoin, α-tocopherol, and interferon-α.66 Nevertheless, folate is characterized by a lower grade of toxicity and our results are encouraging because we obtained a complete response rate of 27% and never observed any clinical or histologic progression during a 6-month treatment.

The correlation of the histologic score with the response score suggests that dysplastic, more risky lesions are more responsive to folate supplementation than inflammatory and hyperplastic ones (Fig. 3). The lower folate levels at baseline and age may be further predictors of responsiveness to folate supplementation, which is helpful for patient selection. In fact, elderly patients with dysplastic leucoplakia had a better response to folate supplementation. Nevertheless, it has been reported that elderly people would be more sensitive to the effect of folate depletion, probably because of a basally impaired DNA repairing machinery.69 These observations are consistent with the hypothesis that folate insufficiency is a long-term risk factor, which over the years increases the rate of carcinogenic progression after exposure to environmental carcinogens by impairing DNA repairing processes.51 In fact, folate is a fundamental cofactor for the synthesis of dNTPs, a substrate of both DNA-synthesizing enzymes and of DNA repairing-enzymes.21, 47

Such results open intriguing perspectives, considering also the fact that hypofolatemia has been reported in 2001 to be the most frequent vitamin deficiency in the US population.52 Folate supplementation, alone or in combination with other chemopreventive drugs, could effectively reduce the risk of progression in an already genetically altered mucosa, especially in patients with hypofolatemia. Folic acid can be simply and safely administered for a long time at a dosage of 10-15 mg a day because it is a virtually nontoxic compound with good oral bioavailability.

Therefore, we are now planning a wide randomized, double-blind multicentric clinical trial to evaluate the effectiveness of long-term folate supplementation in the chemoprevention of HNSCC after endoscopic resection of a laryngeal leucoplakia. In fact, the main issue in the management of laryngeal leucoplakia is not its resection, which is almost always quite comfortable, but the tendency to recur, often with a histologic progression and even transformation.55, 65


  1. Top of page
  2. Abstract