SciELO - Scientific Electronic Library Online

Home Pagealphabetic serial listing  

Services on Demand



  • text new page (beta)
  • English (pdf)
  • Article in xml format
  • How to cite this article
  • SciELO Analytics
  • Curriculum ScienTI
  • Automatic translation


Related links


Arquivos de Gastroenterologia

Print version ISSN 0004-2803On-line version ISSN 1678-4219

Arq. Gastroenterol. vol.53 no.1 São Paulo Jan./Mar. 2016 



Francisco TUSTUMI1 

Flavio Roberto TAKEDA1 

Cintia Mayumi Sakurai KIMURA1 

Rubens Antônio Aissar SALLUM1 



1Departamento de Cirurgia do Aparelho Digestivo do Hospital das Clínicas de São Paulo; Faculdade de Medicina da Universidade de São Paulo, São Paulo, SP, Brasil.



Esophageal cancer is one of the leading causes of mortality among the neoplasms that affect the gastrointestinal tract. There are several factors that contribute for development of an epidemiological esophageal cancer profile in a population.


This study aims to describe both clinically and epidemiologically the population of patients with diagnosis of esophageal cancer treated in a quaternary attention institute for cancer from January, 2009 to December, 2011, in Sao Paulo, Brazil.


The charts of all patients diagnosed with esophageal cancer from January, 2009, to December, 2011, in a Sao Paulo (Brazil) quaternary oncology institute were retrospectively reviewed.


Squamous cell cancer made up to 80% of the cases of esophageal cancer. Average age at diagnosis was 60.66 years old for esophageal adenocarcinoma and 62 for squamous cell cancer, average time from the beginning of symptoms to the diagnosis was 3.52 months for esophageal adenocarcinoma and 4.2 months for squamous cell cancer. Average time for initiating treatment when esophageal cancer is diagnosed was 4 months for esophageal adenocarcinoma and 4.42 months for squamous cell cancer. There was a clear association between squamous cell cancer and head and neck cancers, as well as certain habits, such as smoking and alcoholism, while adenocarcinoma cancer showed more association with gastric cancer and gastroesophageal reflux disease. Tumoral bleeding and pneumonia were the main causes of death. No difference in survival rate was noted between the two groups.


Adenocarcinoma and squamous cell carcinoma are different diseases, but both are diagnosed in advanced stages in Brazil, compromising the patients' possibilities of cure.

HEADINGS Adenocarcinoma; Squamous cell carcinoma; Epidemiologic factors; Period analysis



Câncer esofágico é uma das principais causas de morte por câncer dentre as neoplasias do trato gastrointestinal. Há diversos fatores que contribuem para o desenvolvimento de um perfil epidemiológico de câncer de esôfago em uma população.


Este estudo visa descrever tanto clínica quanto epidemiologicamente a população de pacientes com diagnóstico de câncer esofágico tratados em um instituto quaternário de atendimento ao câncer desde janeiro de 2009 a dezembro de 2011, em São Paulo, Brasil.


Os prontuários de todos os pacientes diagnosticados com câncer de esôfago de janeiro de 2009 a dezembro de 2011 em um Instituto quaternário de tratamento oncológico foram revisados retrospectivamente.


Carcinoma epidermóide foi responsável por 80% dos diagnósticos de câncer esofágico. Idade média ao diagnóstico foi de 60 anos para adenocarcinoma (EA) e 62 para carcinoma epidermóide e o tempo médio entre início dos sintomas até o diagnóstico foi de 3,52 meses para adenocarcinoma e 4,2 para carcinoma epidermóide. O tempo médio para iniciar tratamento foi de 4 meses para adenocarcinoma e 4,42 meses para carcinoma epidermóide. Houve uma clara associação entre carcinoma epidermóide e neoplasias de cabeça e pescoço, bem como com alguns hábitos, tais como tabagismo e etilismo. Adenocarcinoma, por sua vez, mostrou-se associado a câncer gástrico e doença do refluxo gastroesofágico. Sangramento tumoral e pneumonia foram as principais causas de morte para ambos os tipos de câncer. Não foi observada diferença na sobrevida entre os dois grupos.


Adenocarcinoma e carcinoma epidermóide são doenças diferentes, mas ambas ainda são diagnosticadas em estados avançados no Brasil, comprometendo a possibilidade de cura dos pacientes.

DESCRITORES Adenocarcinoma; Carcinoma de células escamosas; Fatores epidemiológicos; Análise transversal


Esophageal cancer (EC) is among the ten most common cancer worldwide and it is one of the leading causes of mortality among the neoplasms that affect the gastrointestinal tract. During the last four decades, the incidence of EC increased significantly8. In Brazil, for instance, it is the 6th most frequent cancer in men and 15th in women. In the year of 2014, 10,780 new cases of EC were diagnosed in Brazil and, in 2011, it was responsible for 7,636 deaths. Even though the survival rate for patients with EC has improved over the decades, it is still very poor, for curative treatment is only destined to patients with early or, at most, locally advanced disease6.

The mortality and incidence trends for EC have varied in many countries over the last decades, especially regarding histological subtypes - adenocarcinoma and squamous cell carcinoma. There are marked differences between these subtypes in terms of incidence, natural history and treatment outcomes. Over the world, but most importantly in Western Europe and North America, it has been observed a higher incidence of esophageal adenocarcinoma (EA), in comparison with squamous cell carcinoma (SCC), which is possibly due to the increase of obesity and gastroesophageal reflux (GERD) incidence in those regions. Even so, SCC still accounts for about 90% of cases of EC worldwide. In Brazil, where SCC is also predominant, corresponding to 96% of all EC cases, it has also been noticed an increasing incidence of esophageal adenocarcinoma7 12 14.

An important characteristic of esophageal cancer is its strong association with known risk factors, for example alcoholism, tabagism, Barrett esophagus (BE) and human papiloma virus infection. Also, a marked characteristic of EC is its heterogenic incidence, higher in Asian countries and a milder incidence in European and American countries. A great diversity is also observed in Brazil, with four cases in every 100 thousand inhabitants in the North of the country and 15 cases for the same size population in the South.

Given the important participation of esophageal cancer in incidence and mortality rates and the observed changes regarding the prevalence of EA, this study aims to describe both clinically and epidemiologically the population of patients with diagnosis of EC treated in a quaternary attention institute for cancer from January, 2009 to December, 2011, in Sao Paulo, Brazil, in order to establish a profile of these patients and compare it with data of other services.


The charts of all patients diagnosed with esophageal cancer from January, 2009, to December, 2011, in a Sao Paulo (Brazil) quaternary oncology institute (ICESP - Instituto do Cancer do Estado de Sao Paulo) were retrospectively reviewed.

Patients were analyzed according to demographic aspects (gender, age, age of diagnosis, body mass index, ethnicity and performance status), known risk factors (alcoholism, smoking, Barrett esophagus, achalasia), associated diseases and neoplasms, localization of the tumor, histologic type, staging, treatment, complications and survival.

The population studied was composed of 565 individuals (n=565), of which 20.7% were female and 79.3% male. The age at diagnosis ranged from 32 to 94 years old.

Regarding statistical analysis, for contingency tables, a Fisher exact test was used; to compare independent means, Student-t test was empolyed.


Among all patients (n=565), 105 (19.13%) had the diagnosis of adenocarcinoma, 444 (80.87%) had squamous cell carcinoma, and 16 (2.83%) had other esophageal neoplasms, such as neuroendocrine or gastrointestinal stromal tumors.

The mean time between first symptoms and diagnostic was 4.31±2.85 months and the mean time to initiate treatment was 4.01±3.12 months.

Comparing gender prevalence between EA and SCC, there were no significant differences - both affect predominantly males, who represented 78.4% of patients with EA and 82.9% of patients with SCC ( P =0.351).

Racial distribution of incidence is illustrated in Figures 1 and 2, both showing a prevalence of Caucasian patients in both subtypes of EC. Statistical analysis of this distribution showed that EA affects Caucasians in a significantly higher proportion than black, Asian or grayish brown patients ( P=0.01).

FIGURE 1 Racial distribution of Esophageal Adenocarcinoma 

FIGURE 2 Racial distribution of Esophageal Squamous Cell Carcinoma 

Average age at the diagnosis was 60.66 years old (StDev 9.73 years old) for SCC and 62.46 years old (StDev 13.03 years old) for EA, with no statistic difference (P =0.213). Average time from the beginning of symptoms to the diagnosis was 3.52 (StDev 2.39) months for EA and 4.2 (StDev 3.04) months for SCC (P -value for the difference=0.036). Average time for initiating treatment when esophageal cancer is diagnosed is 4 (StDev 2.42) months for EA and 4.42 (StDev 2.96) months for SCC ( P -value for the difference=0.22).

Initial clinical stage is seen in Figures 3 and4, showing a predominance of advanced stages at diagnosis (IIIC and IV) in more than half of the cases, in both EA and SCC.

FIGURE 3 Initial clinical stage in esophageal adenocarcinoma 

FIGURE 4 Initial clinical stage in esophageal squamous cell carcinoma 

Body mass index (BMI) at the time of diagnosis and weight loss after beginning of symptoms are described in Table 1. Adenocarcinoma patients had a higher BMI at diagnosis, but SCC patients had a higher weight loss.

TABLE 1 Body mass index (BMI) and weight loss after begin of symptoms. Adenocarcinoma patients had a higher BMI at diagnosis, but squamous cell carcinoma patients had a higher weight loss 

BMI and weight loss Adenocarcinoma Squamous cell carcinoma 95% CI for difference
BMI at the diagnosis (kg/m²) 23.31 20.43 1.603; 4.151
BMI variation since initial symptoms to the diagnosis (kg/m²) 3.97 3.87 -0.847; 1.038
BMI variation per month (kg/m² per month) 1.14 3.87 -3.282; -2.174
Weight loss (%) 14.6 16.6 -5.25; 1.39

Status performance at the diagnosis was evaluated by ECOG and Karnofsky (KPS) functional evaluation, with no significant difference of performance status among patients with EA and SCC.

Regarding associated neoplasms, 18.4% of patients diagnosed with EA had an associated neoplasm - previous, posterior or synchronic to the EC, 42.1% of which were synchronic. For SCC, however, only 12.4% had associated neoplasm, 42.6% of which were synchronic. As seen in Table 2, the neoplasms associated with SCC were predominantly head and neck tumors.

TABLE 2 Main associated neoplasms. Head and neck tumors are much more often found in association to squamous cell carcinoma patients, meanwhile gastric adenocarcinoma is more often associated with gastric adenocarcinoma 

Main associated neoplasms Neoplasms associated with SCC (%) Neoplasms associated with EA (%) P-value for the difference
Colorrectal neoplasm 0 3 0.101
Head and neck SCC 7 1 0.005
Gastric adenocarcinoma 0 5 0.020
Prostatic cancer 0 1 0.476

SCC: squamous cell carcinoma; EA: esophageal adenocarcinoma.

Familiar oncologic past was also investigated (see Table 3), but only family history of colorrectal cancer was statistically associated with higher incidence in EA.

TABLE 3 Familial oncological past. History of colorrectal tumors in the family is more common in esophageal adenocarcinoma patients 

Familial oncological past EA- At least one first degree family member with neoplasm (%) SCC- At least one first degree family member with neoplasm (%) P-value for the difference
Colorrectal cancer 8 2 0.031
Gasric adenocarcinoma 6 3 0.203
Esophageal cancer 3 2 0.100
Prostate cancer 5 2 0.353
Breast cancer 7 4 0.227
Head and neck cancer 3 5 0.386
Cervical cancer 3 1 0.246
Lung cancer 6 3 0.172

Habits and addictions were also evaluated. Alcoholism and smoking were more prevalent among SCC patients in comparison with EA patients ( P <0.05). The mean smoking load (packs per day multiplied by years smoking) was 20.89 (SE mean 3.4) for EA patients and 45.27 (SE mean 3.96) for SCC patients.

Presence of known association with non-oncologic diseases is described in Table 4, showing that Barrett esophagus and Gastroesophageal Reflux Disease (GERD) are more frequent in EA patients. On the other hand, chronic obstructive pulmonary disease (COPD) is more common in SCC patients.

TABLE 4 Comorbidities. Barrett esophagus and GERD are more frequent in esophageal adenocarcinoma patients. On the other hand, chronic obstructive pulmonary disease (*COPD) is more common in SCC patients 

Non-oncologic associated diseases EA (%) SCC (%) P-value for the difference
Diabetes mellitus 15 9 0.062
Systemic hypertension 34 28 0.167
Obesity 8.4 3.58 0.058
Barrett esophagus 12 0 0.000
GERD 17 0 0.000
Hiatal hernia 9 0 0.001
Pulmonary tuberculosis 3 5 0.076
Chagas disease 1 2 0.289
COPD* 4 9 0.021

EA: esophageal adenocarcinoma; SCC: squamous cell carcinoma.

The survival rate of patients with EA was of 13.2 months (±StDev 10.2 months) and of 14 months (±StDev 11.2 months) for patients with SCC, with no significant difference between them.

The main causes of death and main complications associated with EC were also studied and are described in Tables 5 and 6, respectively. Tumoral bleeding and pneumonia were the leading causes of death for both EA and SCC patients, even though tumoral bleeding, tracheoesophageal fistula and pulmonary abscess/pleura empyema are far more often in SCC.

TABLE 5 Causes of death. There was no statistical significance in causes of death for esophageal adenocarcinoma or squamous cell carcinoma 

Causes of death EA (%) SCC (%)
Pneumonia 42 64
Tumoral bleeding 14 11
Massive hemoptysis 2 1
Pulmonary thromboembolism 2 1
Acute abdominal obstruction 6 0
Vascular acute abdomen 2 0
Cerebrovascular accident 2 0
Cholangitis 2 0
Acute pulmonary edema 2 0
Infectious endocarditis 4 0
Surgical complications 2 1
Intracranial hypertension 2 1
Urinary tract infection 4 2
Acute renal injury 4 3
Febrile neutropenia 4 0
Hepatorenal syndrome 2 0
Other causes of sepsis (excluded pneumonia and urosepsis) 14 11

EA: esophageal adenocarcinoma; SCC: squamous cell carcinoma.

TABLE 6 Complications in esophageal carcinoma. Tumoral bleeding, tracheoesophageal fistula and pulmonary abscess/pleura empyema are far more often in squamous cell carcinoma, meanwhile venous or pulmonary thromboembolism is more common in esophageal adenocarcinoma 

Complications EA (%) SCC (%) P-value for the difference
Tumoral bleeding 6 13 0.044
Tracheoesophageal fistula 0 10 0.000
Venous or pulmonary thromboembolism 9.52 3 0.022
Pulmonary abscess or pleural empyema 0 4.95 0.003

EA: esophageal adenocarcinoma; SCC: squamous cell carcinoma.


Incidence rates of esophageal cancer are very heterogeneous worldwide and even in the same country. Southern Brazil, for instance, has the highest rate of esophageal cancer in South America, meanwhile Northern Brazil has a very low incidence of the disease (4:100 000 inhabitants)2 12.

Even though many studies show a changing pattern in the incidence of esophageal adenocarcinoma when compared to squamous cell carcinoma worldwide and in São Paulo2 9, in our service, there is still an absolute predominance of the latter. A recent research performed in Rio Grande do Sul State (Southern Brazil), showed that not only the prevalence rate of esophageal cancer, along a 20 year-period, remained stable, but also, the proportions of SCC and ADC did not change- 92% of the cases are SCC and 7.9% are EA5.

Some studies also show a higher prevalence of esophageal carcinoma in black males, with a trend of increasing incidence of adenocarcinoma among white males. In our service, however, most of the patients with EC, either EA or SCC, were white males, and the predominance of white patients in the population with EA was statistically significant ( P =0.001)2.

The association between SCC and smoking and alcoholism found in this study is consistent with literature11. Also, the association of SCC with head and neck tumors, found in the present study, is well reported in other researches and is probably related to the fact that exposition to tobacco and alcohol is present in most patients with those diseases.

An interesting finding of this study, which does not correlate with current worldwide data is the higher incidence of EC among white males. Most of the information available today shows a higher incidence among afro-descendants. However, in this study, the vast majority of patients were male Caucasians, with statistically relevance in patients with EA, in special. This may be due to the higher number of white people in Sao Paulo than other Brazilian places3 13.

Regarding associated neoplasms, meanwhile SCC is known to be frequently associated with oropharingeal tumors (probably because of field cancerization), it seems that EA is associated with gastric adenocarcinoma - the incidence of gastric adenocarcinoma in patients with EA was 500:10 000 and in the general Brazilian population it is about 1:10 000. This finding is also in accordance with literature4 10.

In this study, familial history of esophageal cancer did not correlate to increased risk of EC. However, familial history of colorrectal cancer was associated with higher risk of esophageal adenocarcinoma. No conclusive studies about familial past and esophageal cancer have been performed so far. There are some researches, though, that suggest possible association between Barrett esophagus with colorrectal cancer, but the mechanism by which this occurs is not well established yet1 15.

Another important finding of this study is the long time for diagnosing the cancer. It was noted that the time between first symptoms and diagnosis is even longer for SCC. That leads to more advanced stages at diagnosis, and, added to another long period until starting treatment in Brazil, hinders the patients' chances of curative treatment.


Each population has its demographic features in esophageal cancer, due to each particularity in population habits and genetics. Adenocarcinoma and squamous cell carcinoma are indeed different diseases, from epidemiological features to biological behavior. Nevertheless, Brazilian late diagnosis and delay in starting treatment are factors that contribute to the high incidence of advanced stages for both diseases. More effort should be made in order to earlier diagnose and initiate treatment of these patients so that their prognosis can be improved.


Andrici J, Tio M, Cox MR, Eslick GD. Meta-analysis: Barrett's Oesophagus and the Risk of Colonic Tumours. Aliment Pharmacol Ther. 2013;37:401-10. [ Links ]

Devesa SS, Blot WJ, Fraumeni JF Jr. Changing patterns in the incidence of esophageal and gastric carcinoma in the United State. Cancer. 1998;83:2049-53. [ Links ]

Dominitz JA, Maynard C, Billingsley BG, Boyko DJ. Race, treatment, and survival of veterans with cancer of the distal esophagus and gastric cardia. Med Care. 2002;(1 Suppl): I14-26. [ Links ]

Durazzo MD, Araujo CEN, Neto B, Potenza AS, Costa P, Takeda Fl et al. Clinical And Epidemiological Features Of Oral Cancer In A Medical School Teaching Hospital From 1994 To 2002: Increasing Incidence In Women, Predominance Of Advanced Local Disease, And Low Incidence Of Neck Metastases. Clinics. 2005;60:293. [ Links ]

Fagundes RB, de Carli D, Xaubet RV, Cantarelli JC Jr. Unchanging pattern of prevalence of esophageal cancer, overall and by histological subtype, in the endoscopy service of the main referral hospital in the central region of Rio Grande do Sul State, in Southern Brazil. Dis Esophagus. 2015. doi: 10.1111/dote.12350 [Epub ahead of print] [ Links ]

Ferlay J, Soerjomataram I, Ervik M, et al. GLOBOCAN 2012 v1.0, Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 11 [Internet]. Lyon, France: International Agency for Research on Cancer; 2013. Available from: [ Links ]

Instituto Nacional de Cancer [Internet]. Incidência de Câncer no Brasil (BR); 2014. Available from ]

Jemal A, Siegel R, Ward E, et al: Cancer statistics. CA Cancer J Clin. 2009,59:225-49. [ Links ]

Kroep S, Lansdorp-Vogelaar I, Rubenstein JH, Lemmens VE, van Heijningen EB, Aragonés N, van Ballegooijen M, Inadomi JM. Comparing trends in esophageal adenocarcinoma incidence and lifestyle factors between the United States, Spain, and the Netherlands. Am J Gastroenterol. 2014;109:336-43. [ Links ]

Lee GD, Kim YH, Kim JB, Choi SH, Kim HR, Kim DK, Park SI. Esophageal Cancer Associated with Multiple Primary Cancers: Surgical Approaches and Long-term Survival. Ann Surg Oncol. 2013;20:4260-6. [ Links ]

Pickens A, Orringer MB. Geographical Distribution and Racial Disparity in Esophageal Cancer. Ann. Thorac Surg. 2003;76:S1367-9. [ Links ]

Rustgi AK, El-Serag HB. Esophageal Carcinoma. New England Journal of Medicine, 2014; 371:2499-509. [ Links ]

Siewert JR, Ott K. Are Squamous and Adenocarcinomas of the Esophagus the Same Disease? Semin Radiat Oncol. 2007;17:38-44. [ Links ]

Tercioti-Junior Valdir, Lopes Luiz Roberto, Coelho-Neto João de Souza. Adenocarcinoma versus carcinoma epidermóide: análise de 306 pacientes em hospital universitário. ABCD, Arq Bras Cir Dig. 2011;24:272-6. [ Links ]

Vaughan TL, Kiemeney LA, McKnight B. Colorectal cancer in patients with esophageal adenocarcinoma. Cancer Epidemiol Biomarkers Prev. 1995;4:93-7. [ Links ]

3Disclosure of funding: no funding received

Received: June 17, 2015; Accepted: September 12, 2015

Correspondence: Flavio Roberto Takeda. Rua Capote Valente, 640, ap. 172 - Pinheiros - CEP: 05409-002 - São Paulo, SP, Brasil.

Authors' contributions Kimura CMS: data collection and text composing. Tustumi F: data collection and satistical analysis. Takeda FR: research design and execution. Sallum RAA: research design. Ribeiro Junior U: research design. Cecconello I: research design.

Declared conflict of interest of all authors: none

Creative Commons License This is an open-access article distributed under the terms of the Creative Commons Attribution License