Colorectal carcinoma in poor blacks


  • Harold P. Freeman M.D.,

    Corresponding author
    1. Department of Surgery, Harlem Hospital Center, College of Physician and Surgeons of Columbia University, New York, New York
    • Department of Surgery, North General Hospital, 1879 Madison Avenue, New York, NY 10035
    Search for more papers by this author
    • Fax: (212) 423-1594

  • Tarek A. Alshafie M.D.

    1. Department of Surgery, Harlem Hospital Center, College of Physician and Surgeons of Columbia University, New York, New York
    Search for more papers by this author



Death rates for most cancers continue to be higher for African Americans, particularly those in inner cities. Harlem Hospital serves a poor, predominantly African-American community in New York City.


Tumor registry records for 615 patients treated for colorectal carcinoma at Harlem Hospital between 1973 and 1992 were reviewed.


Of the patients, 45.2% were male and 54.8% female, 97.2% were black, and 82% resided in Harlem. All patients were symptomatic at the time of diagnosis;15.3% were first diagnosed intraoperatively; 8.4% were in American Joint Committee on Cancer Stage I, 20.8% Stage II, 22.8% Stage III, 39.0% Stage IV, and 8.0% could not be staged. Colon resection with intention of cure was performed on 50.6%, 21.5% had palliative resection, and 11.6% had colostomy or other palliative surgery. Adjuvant chemotherapy or radiotherapy was given to 6.2%; 16.9% had no surgical treatment because of advanced stage, poor condition, or refusal of surgery; 12.7% presented with perforation or intestinal obstruction. Operative mortality was 15.3% overall and 10.6% for 311 patients who had surgery with intention of cure. Twenty-five patients had local recurrence, 86 had subsequent distant metastases, and 33 patients had both local and distant recurrence. Forty-nine patients (8%) were lost to follow-up. The 5-year crude survival rate for 615 patients was 18.7%. The relative survival rate was 19.7%, substantially lower than the national average for the same years.


Although colorectal carcinoma mortality continues to decline nationally, in this population of poor blacks the mortality rate remained high and unchanged. The most important cause of this is late presentation at an incurable stage, resulting from the combined effects of poverty, lack of education, and lack of access to primary care. Culturally sensitive educational programs and accessible health care systems for the poor are needed. Cancer 2002;94:2327–32. © 2002 American Cancer Society.

DOI 10.1002/cncr.10486

Considering men and women together, colorectal carcinoma is the most common cancer and the second leading cause of cancer death in the United States.1 Approximately 7% of Americans will develop colorectal carcinoma within their lifetime.2 Colorectal carcinoma incidence increased in the United States from 1973 through 1985 and then decreased through 1995. The U.S. mortality rate from colorectal carcinoma has declined among women since 1950 and among men since 1985. At all ages, men are more likely to develop and die from colorectal carcinoma than women.3 African Americans have higher colorectal carcinoma incidence and mortality rate compared with all other ethnic groups.4 Between 1960 and 1992, colorectal carcinoma mortality increased 53% among African-American males and 8% among African-American females.5

This report is a retrospective study of 615 patients treated at an inner-city public hospital over a 20-year period between 1973 and 1992. Ninety-seven percent of the patients were black, and virtually all were poor.



Six hundred fifteen consecutive patients were diagnosed with colorectal carcinoma, were treated at Harlem Hospital Center, and were followed up by the hospital tumor registry during a 20-year period between January 1973 and December 1992. Eighty-two percent were permanent residents in the Harlem community, and 97.2% were black. Central Harlem is a neighborhood in upper Manhattan just north of Central Park. Its population is 96% black and has been predominantly black since before World War I.6 The families of 41% of the people of Harlem had incomes below the government-defined poverty line in the 1980 and 1990 censuses. Patients whose disease was diagnosed and treated elsewhere and later admitted to Harlem Hospital (usually in a terminal condition) were excluded from the study. Less than 3% of patients examined were excluded on this basis.

Registry Information

Tumor registry records were reviewed to determine the age, gender, race, presenting symptoms, associated illness, family history of cancer, pathologic staging of the tumor, treatment received, recurrence after treatment, and status after 5 years. To facilitate comparisons with other reports the patients were classified in three different ways: according to the TNM staging (American Joint Committee on Cancer [AJCC] protocol for staging of colorectal carcinoma),7 Dukes staging,8 and the SEER staging categories.9

Statistical Methods

The data were analyzed with SAS statistical software programs. Survival times were obtained using the actuarial life table method.10 To adjust the survival rates for deaths from other causes than colorectal carcinoma, and to allow for comparison of our rates with other databases, we calculated relative survival rates (RSR) as the ratio of the observed survival rate (OSR) to the expected survival rate (ESR).11 Observed survival rates were calculated by the life table method, and the ESR was obtained for a group of people in the general population in the State of New York during the same period of time, who were similar with respect to age, gender, and race.5 The standard error (SE) of the OSR was computed from Greenwood formula12 and the SE of RSR was computed as SE of (OER)/ESR.

Use of Race as a Variable

The authors' assumptions are that racial categories as used in this study are socially and politically determined. We do not attribute any biologic meaning to the racial categories, white and black.


Age, Race, and Gender

Of the 615 patients identified, 278 patients (45.2%) were male, and 337 (54.8%) were female. The age range was from 18 to 96 years with a mean of 68.1 years; 34.8% were younger than 65 years. Five hundred ninety-eight patients (97.2%) were black, 10 Hispanic, and 6 white and 1 patient was of Asian descent (Table 1).

Table 1. Patients with Colorectal Carcinoma in Harlem Hospital Center
CharacteristicAll (%)Male (%)Female (%)
Gender615 (100%)278 (45.2%)337 (54.8%)
Age (yrs)
 < 3514 (2.3)86
 35–4410 (1.6)55
 45–5455 (8.9)2728
 55–64135 (22)6570
 65–74203 (33)96107
 ≥75198 (32.2)105121
 <65214 (34.8)105109
 ≥65401 (65.2)173228
 Black598 (97.2)
 Hispanic10 (1.6)
 White6 (1.0)
 Asian1 (0.2)


All patients had symptoms at the time of presentation; 38.8% presented with bleeding per rectum, 32.2% with abdominal pain, 21.1% with bowel obstruction, 17.8% with weight loss, and 17.8% with changes in the bowel habits. Symptoms caused by locally advanced disease and/or distant metastasis, e.g., perforation, peritonitis, abdominal mass, jaundice, and anemia, were the main presentation in 24.4% (Table 2).

Table 2. Presenting Symptoms
SymptomNo. of patients (%)
  • a

    Perforation, peritonitis, abdominal mass, jaundice, severe anemia.

Bleeding per rectum239 (38.8)
Abdominal pain198 (32.2)
Bowel obstruction130 (21.1)
Change in bowel habits109 (17.8)
Weight loss109 (17.8)
Others(a)150 (24.4)

Methods of Diagnosis

Several methods of diagnosis were used. Many underwent more than one method. Endoscopy was the main method of diagnosis in 380 patients (61.8%), barium enema in 206 (33.5%). Computed tomography or liver scan was used in 106 patients (16.7%). In 94 patients (15.3%), the diagnosis was first made at the time of surgery.

Site of Tumor

In 143 patients (23.3%), the tumor was located in the ascending colon, 92 were (14.9%) in the transverse colon, 53 (8.6%) had cancer in the descending colon, 159 (25.9%) had cancer in the sigmoid colon, 38 (6.2%) had cancer in the rectosigmoid region, and 130 (21.1%) had rectal carcinoma. We analyzed the patients according to 5-year intervals to see whether there was any shift in the site of tumors over time. Our results showed no significant difference in the site of the tumor over the 20-year period (Table 3).

Table 3. The Site of Colorectal Carcinoma in Harlem Hospital
SiteAll Years (%)1973–1977 (%)1978–1982 (%)1983–1987 (%)1988–1992 (%)
Ascending143 (23.3)37 (23)48 (26.8)38 (24.4)20 (16.8)
Transverse92 (14.9)25 (15)28 (15.6)21 (13.5)18 (15.1)
Descending53 (8.6)8 (5)14 (7.8)20 (12.8)11 (9.2)
Sigmoid159 (25.9)45 (28)45 (25.2)34 (21.6)35 (29.4)
Rectosigmoid38 (6.2)11 (6.6)13 (7.3)8 (5.1)6 (5)
Rectal130 (21.1)35 (21.7)31 (13.3)35 (22.4)29 (24.5)


Considering all patients at the time of presentation, according to the AJCC classification, 6 had carcinoma in situ, and 52 (8.4%) were in Stage I, 128 (20.8%) in Stage II, 140 (22.8%) in Stage III, 240 (39%) in Stage IV, and 49 were unstaged.

Sixty-five patients (10.6%) were in Dukes A; 170 (27.6%) in Dukes B; 254 (41.3 %) in Dukes C. One hundred twenty-six (20.5%) were too sick to have an operation, refused surgery, or died before treatment could be administered; therefore, these patients were unstaged.

One hundred eighty-six (30.2%) had localized disease, 140 (22.8%) had regional spread, and 240 (39%) were found to have distant metastases. The distribution of stages according to the three classification schemes is shown in Table 4. The distribution of stage at presentation for the most recent 5 years of this study is compared with SEER data for the same time in (Fig. 1).

Table 4. Stage of The Tumor at Presentation
  1. AJCC: American Joint Committee on Cancer; X: unstaged.

 Regional spread14022.8
 Distant spread24039
Figure 1.

Stage distribution at presentation. Harlem: Harlem Hospital Center (1986–1992); U.S. white and U.S. black: SEER data (1986–1992).

Patient Management

Four hundred forty-three patients (72%) underwent colon resection, 311 (50.6%) of them with intention of cure. Sixty-eight (11.6%) had other palliative surgery. Eighteen (2.9%) had adjuvant chemotherapy, 15 (2.4%) had adjuvant radiotherapy, and 5 had both chemotherapy and radiotherapy adjunct to the surgical treatment. Twelve patients had no surgical intervention but received other treatment modalities: 3 had only chemotherapy, 7 only radiotherapy, and 2 received both. Ninety-two (15%) had no treatment because of poor condition, advanced stage, or refusal of surgery.

Of 511 patients who underwent surgery, 78 (15.3%) died in the postoperative period. Six of them were in Stage I, 11 in Stage II, 16 in Stage III, 41 in Stage IV, and 4 were unstaged.

Of 311 who underwent surgery with the intention of cure, 2 of them were in situ, and 52 were in Stage I, 128 were in Stage II, and 129 were in Stage III. Twenty-five of these patients (8.4%) had local recurrence, 86 (27.6%) had recurrence with distant metastases, and 33 (10.6%) had both local recurrence and distant metastasis. For treatment of recurrence, 32 underwent surgery, 22 had radiotherapy, and 17 had chemotherapy.

Five hundred sixty-six patients (92%) had follow-up for 5 years or until they died. Forty-nine (8%) were lost in the follow-up. Seven of those lost in the follow-up were in Stage I, 13 were in Stage II, 14 were in Stage III, 2 were in Stage IV, and 13 were unstaged. Seventeen of them were younger than 65 years of age, and 32 were 65 years and older.

The crude 5-year survival for 615 patients was 18.7%. The 5-year survival corrected for those lost to follow-up was 20.3%. The corrected 5-year survival for the patients younger than 65 years of age was 28.9% (57 patients), whereas the corrected 5-year survival for those 65 years and older was 15.7% (58 patients).

The corrected 5-year survival according to the site of the tumor was 19.6% for tumor of the cecum and the ascending colon, 15.2% for transverse colon, 25% for the descending colon, 19% for the sigmoid colon, and 16.1% for the rectum (including rectosigmoid tumors).

Corrected 5-year survival according to the stage of the disease at presentation was 64.1% for Stage I, 48.1% for Stage II, 25.5% for Stage III, and 28.1% for the unstaged patients.

Corrected 5-year survival for the patients younger than 65 years of age was 31.6%, and for those 65 years and older it was 18.5%. If the postoperative deaths were excluded, the overall 5-year survival rate was 23.5%.

The 5-year RSRs presented by the year of diagnosis and age at diagnosis are shown in Table 5. The differences between these rates and the rates of the SEER report9 are statistically significant (P ≤ 0.05). For all time intervals from 1974 to 1992, the 5-year relative survival rate was much worse than the national rates for white or black colorectal carcinoma patients (Fig. 2.) The national results are improving whereas survival in the study population remains low and essentially unchanged.

Table 5. 5-Year Relative Survival Rate (%)
Years of DiagnosisHarlemSEER (white)SEER (blacks)
Figure 2.

Five-year survival rate. Harlem: Harlem Hospital Center; U.S. white and U.S. black: SEER data (1974–1992).


In a previous report, we showed that in recent decades overall mortality rates have declined for both white and nonwhite Americans, but national averages obscure the extremely high mortality rates in inner-city communities. We concluded that Harlem and other similar areas with largely black populations have mortality rates so high as to justify special consideration analogous to that given to natural-disaster areas.6 We also have demonstrated that African Americans, particularly inner-city African Americans, suffer higher cancer mortality from various causes than the mainstream American society.13, 14

In this study, the most common conditions with increased risk for colorectal carcinoma, first-degree family history of cancer and personal history of malignancy, were comparable to other reports.15–17 Several factors probably contributed to the very low relative 5-year survival rate in our patients.

Previous studies show that the stage of the disease at the time of diagnosis is the most important factor in the prognosis of colorectal carcinoma, and timely excision provides the only realistic chance of cure.18, 19 In comparisons with previous reports, the patients in this study presented with more advanced disease (Table 4). In most previous reports, 40– 50% of cases were limited to the bowel wall, 30– 33% of lesions had regional lymph node metastasis at the time of diagnosis, and 25% had distant metastatic lesions.15, 20–25 We found 30.2% limited to the bowel wall, 22.8% with only regional lymph node spread, and 39% with distant metastases. The proportion with distant metastases is significantly higher than that reported by SEER for all patients, for whites and for blacks (Fig. 1).

Recent reports have shown a trend toward early diagnosis of colorectal carcinoma in asymptomatic patients.26, 27 Beart et al. report that among proximal colon carcinoma patients the proportion who were symptomatic at diagnosis decreased from 90% to 77% between 1940–1959 and 1960–1979. The proportion of sigmoid carcinoma patients who were symptomatic declined from 94% in 1940–1959 to 82% in 1960–1979.27 Ahlquist et al. reviewed the records for colorectal carcinoma patients with preceding stool blood test: 29% were asymptomatic.28 In the current study, none of the patients was asymptomatic at the time of the diagnosis.

Even though the most common cause of large bowel obstruction in the United States is colorectal carcinoma, only 14% of patients nationwide with colorectal carcinoma present with a picture of bowel obstruction.16, 29–31 In contrast, 21.1% of the patients in this study had bowel obstruction at diagnosis. Bowel obstruction is known to be an important independent prognostic factor. The overall prognosis is generally worse with this presentation because it is associated with more advanced stage of the disease.16, 30, 31

The distribution of the colon carcinoma in this study (Table 3) was very similar to the National Cancer Data Base report on colorectal carcinoma.32 The trend of proximal migration of colorectal carcinoma reported elsewhere3, 27, 32 was not observed (Table 3). Whether this “proximal migration” is because of evolution in the biology of the colorectal carcinoma or an artifact of more frequent and more extensive access to the proximal bowel via endoscopic approaches remains to be documented.32

Although surgery is considered the most effective therapy for colorectal carcinoma,26, 29 only 311 patients (50.6%) underwent surgery with the intention of cure. Ninety-two patients (15%) had no treatment, much higher than previous reports.21, 32–36 This finding reflects late presentation and poor general condition caused by advanced disease.

The postoperative mortality in the current study, 15.3%, is higher than previous comparable reports that range from 4.5% to 8.5% postoperative mortality.18, 19, 27, 28 Forty-eight of the 78 patients who died in the postoperative period had emergency operations—31 for bowel obstruction and 10 for peritonitis caused by bowel perforation. Umpleby et al. had 36% hospital mortality with patients requiring emergency surgery.30 Kelly et al. reported 17% operative mortality with bowel obstruction and 30% when there was peritonitis caused by perforated tumor.31 Phillips et al. had 23% hospital mortality for large bowel obstruction caused by colon carcinoma.37 Many other reports have shown higher postoperative mortality related to emergency admission..19, 37–39 The unfavorable effects of advanced age and tumor stage on hospital mortality were clearly showed in previous reports,34, 39, 40 as well as our study. The mean age of our patients who died after operation was 72.3 years. Forty-one had Stage IV disease.

SEER data nationwide in the period from 1974 to 1992 indicate that the relative 5-year survival rate improved in all races: from 49.5% to 61.5% overall, and from 44.6% to 52.4 % for blacks. Five-year relative survival rate for all Harlem patients was 19.7%, significantly lower and unchanged during this time interval (Fig. 2).

The 5-year survival rate in Harlem was also lower when compared with previous studies.21, 22–24, 29, 32, 33, 35, 37, 38 The 5-year survival rates for patients with Stage I and II colorectal carcinoma were lower than those in SEER data but similar to those reported in the National Cancer Data Base.32 Increasing evidence suggests that with equal access to the same modalities of treatment there is equal outcome regardless of race in breast,41, 42 prostate,43, 44 cervical,45 and childhood cancer.20 It seems likely that the same will be true for colorectal tumors.


Colorectal carcinoma patients in Harlem Hospital Center have similar age, gender distribution, and location of the tumor in comparison with colorectal carcinoma patients nationwide. They have similar treatment outcome when patients present with potentially curable disease. However, 70% of colorectal carcinoma patients in Harlem Hospital had incurable disease at the time of presentation. In contrast with national findings, all colorectal carcinoma patients in this study had symptoms at the time of the diagnosis. Five-year survival rates were much lower. Although colorectal carcinoma mortality continues to decline nationally, in this population of poor blacks the mortality rate remains extremely high and unchanged. We attribute this primarily to late stage disease at the time of initial treatment. This, we believe, is because of the combined effects of poverty, low education level, and poor or absent primary care, leading to barriers to early diagnosis and treatment. We suggest that addressing the striking difference in mortality in Harlem and in similar populations will require provision of culturally sensitive educational programs combined with health care systems that allow easy access to early detection and treatment of colorectal carcinoma, regardless of the socioeconomic status of the population.