World J. Surg. 28, 716720, 2004 DOI: 10.

1007/s00268-004-7232-8

WORLD
Journal of

SURGERY
2004 by the Socie te Internationale de Chirurgie

Can Early Diagnosis of Symptomatic Colorectal Cancer Improve the Prognosis?

Fernando Gonzalez-Hermoso, M.D., Julian Perez-Palma, M.D., Joaquin Marchena-Gomez, M.D., Nieves Lorenzo-Rocha, M.D., Vincente Medina-Arana, M.D.
Departament of General Surgery, Hospital Universitario Canarias, Ofra s/n. La Cuesta, 38320 La Laguna, Santa Cruz de Tenerife, Spain Published Online: June 16, 2004

Abstract. Patients with colorectal cancer continue to present with relatively advanced tumors. Delay in diagnosis is often believed to have been a contributing factor, and the validity of this hypothesis has seldom been questioned. The aim of this study was to establish whether a delay in diagnosis is related to long-term survival and if the most frequent symptoms were related to the stage or time at which the carcinoma was diagnosed. Data from 660 patients surgically treated for uncomplicated colorectal carcinoma in our institution between 1985 and 2000 were analyzed retrospectively. Age, sex, initial symptoms, duration of symptoms, neoplasm location, curative surgery, TNM stage, and survival time were the variables recorded. Patients were classified into two groups according to symptom duration: < 3 months versus 3 months. Comparative statistical analysis was performed for the two groups as well as the initial symptom, TNM stage, and survival time. Also, the initial symptoms most frequently reported were compared with the TNM stage. The two groups were found to be equal with regard to distribution of age, gender, location of the neoplasm, type of surgery performed, and TNM stage. We found that symptom duration was shortened in the presence of abdominal pain (p = 0.002) [odds ratio (OR) 0.53; 95% confidence interval (CI) 0.350.80] and was delayed in the presence of an anemic syndrome (p = 0.006) (OR 2.4; 95% CI 1.274.56). Also, the stage of the neoplasm was related to rectal bleeding (p < 0.001) and abdominal pain (p = 0.008). The log-rank test indicated that duration of symptoms was not related to long-term survival (p = 0.90). We concluded that the duration of colorectal cancer symptoms is not related to the stage or prognosis of tumors.

studying risk factors. However, it has been suggested that recognition of the disease during the asymptomatic or preclinical period can improve the prognosis [7]. This reinforces the value of colorectal cancer prevention programs because of which adenomas that may eventually prove malignant are identified and excised. We studied a population with colorectal cancer to establish whether symptom duration was related to long-term survival. We also wanted to determine if the distribution or prevalence of the most frequent symptoms was related to the stage or the time at which the carcinoma was diagnosed. Patients and Methods Data from 895 patients diagnosed consecutively and surgically treated for colorectal carcinoma in our institution between January 1985 and December 2000 were analyzed retrospectively. We excluded 139 patients who presented with obstruction or intestinal perforation, another 18 patients whose clinical picture was confused with nonspecific intestinal disease, and 78 patients who were lost during follow-up. Data were collected immediately after the discharge of each patient and supervised by a physician-specialist in coloproctology, who verified the quality of the clinical data. SPSS 9.0 for Windows was used for analysis. The variables recorded from the 660 patients were age, sex, initial symptom that brought the patient to the doctor (change in bowel movements, abdominal pain, rectal bleeding, tenesmus, constitutional syndrome, palpable mass), duration of symptoms, location of the neoplasm, curative surgery versus noncurative surgery, tumor stage according to the TNM system, and survival time expressed in months. Patients were classified into two groups according to symptom duration: a group with symptoms < 3 months versus one with symptoms 3 months. This cutoff point was defined in relation to the median time of evolution of the symptoms in our patient population, which coincides with that of many other studies [813]. Follow-up protocols were used, and in some cases there was periodic consultation by telephone to monitor patient status. The mean follow-up time was 43.19 months (3.60 years). Symptom duration was calculated from the first complaint of the

Patients with colorectal cancer usually complain of one of three clinical pictures: chronic abdominal symptoms, intestinal obstruction, or perforation accompanied by peritonitis. These presentations occur in 77%, 16%, and 7% of patients, respectively [1]. A relation between the duration of symptoms and prognosis has been suggested, although not conclusively [2]. This may be because prognosis depends on multiple factors: clinical course, availability of proper health care, patient attitude, and tumor biology [3]. Prospective studies using multivariate regression analysis [46] have established that the main diagnostic factor is the stage of the disease at the moment of diagnosis, whatever the duration. Only some subgroups of patients are of particular significance when
Correspondence to: Fernando Gonzalez-Hermoso, M.D., e-mail: fgonzale@ull.es

Gonzalez-Hermoso et al.: Colorectal Cancer: Early Diagnosis and Prognosis

717

patient to the date of intervention. Groups of symptoms were not examined, but individual symptoms were. Curative treatment was considered to be surgery that did not leave any macroscopically seen residual neoplasm. Surgery for recurrent neoplasms was not included in the study, nor was preoperative survival. To define the location as the left or right colon, the splenic angle was taken as a reference. The neoplasms were classified according to the TNM system of the American Joint Commission for Cancer (AJCC) [14]. After descriptive analysis of the population studied, a comparative statistical analysis was performed for the duration of symptoms (< 3 months versus 3 months), the age and sex of the patient, neoplasm location, type of surgery (curative versus not curative) initial symptom, and TNM stage. The initial symptoms most frequently reported were compared with the TNM stage. We used Students t-test or the Mann-Whitney U-test to compare continuous variables and the 2 test or Fisher test to compare proportions. The Kaplan-Meier method was used to calculate the main and median survival, and the Table-of-life method was used for graphic representation of survival curves. Lastly, we analyzed the possible association between symptom duration and overall survival time using the log-rank test. A value of p < 0.05 was considered statistically significant. Odds ratio (OR) and 95% confidence interval (95% CI) were also calculated for significant variable associations. Results Of 660 patients, 360 were men (54.5%) and 300 were women (45.5%). The mean age was 66.16 years (95% CI 65.1867.14; SD 12.80; median 67 years), with the range varying from 22 to 94 years. In 313 patients (47.4%), the symptoms were present for 3 months or less, and in 347 patients (52.6%) they were present for more than 3 months at the time of diagnosis. In 136 patients (20.6%), the neoplasm was located on the right side and in 524 patients (79.4%) on the left side. Regarding TNM stage, 99 patients had stage I (15.0%), 266 had stage II (46.3%), 162 had stage III (24.5%), and 133 had stage IV (20.2%) disease. The surgeon was able to perform curative surgery in 499 patients (75.6%) and noncurative surgery in 144 patients (21.8%); the type of surgery was not recorded for 17 patients. The mean survival time according to the Kaplan-Meier method was 91.80 months (SE 3.92; 95% CI 84.1299.48), and the median survival was 66.30 months (SE 6.55; 95% CI 53.4679.14). Based on the duration of symptoms, the two groups were found to be equal with regard to distribution of age, gender, location of the neoplasm, type of surgery performed, and TNM stage (Table 1). Regarding the initial symptomatology (Table 2), the duration of symptoms had a statistically significant relation with the presence of abdominal pain (p = 0.002; OR 0.53; 95% IC 0.350.80), which was more frequent in patients with a symptom duration of < 3 months, and with the existence of an anemic syndrome (p = 0.006) (OR 2.4; 95% IC 1.274.56), which was more frequent in patients with symptoms lasting 3 months. Also, there was a statistically significant relation between the stage of the neoplasm and the most frequently observed symptoms, which were rectal bleeding (p < 0.001) and abdominal pain (p = 0.009) (Table 3). The log-rank test indicated that the duration of symptoms was not related to long-term survival (p = 0.90) (Table 4). There were no significant differences between the two survival curves (Fig. 1).

Table 1. Duration of symptoms in relation to sex, location of neoplasm, character or type of surgery, and TNM stage. Parameter Average age (years) Sex Male Female Neoplasm location Right colon Left colon Type of surgery Curative Noncurative Not available TNM stage I II III IV 3 months 66.27 182 (50.6%) 131 (43.7%) 69 (50.7%) 244 (46.6%) 246 (49.3%) F62 (43.1%) 5 (29.4%) 46 (46.5%) 131 (49.2%) 76 (46.9%) 60 (45.1%) > 3 months 66.06 178 (49.4%) 169 (56.3%) 0.38 67 (49.3%) 280 (53.4%) 0.13 253 (50.7%) 82 (56.9%) 12 (70.6%) 0.87 53 (53.5%) 135 (50.8%) 86 (53.1%) 73 (54.9%) p 0.83 0.08

Table 2. Relation between the initial symptoms and the duration of symptoms. Parameter Change in bowel habit Yes No Abdominal pain No Yes Rectal bleeding No Yes Anemia No Yes Tenesmus No Yes Abdominal mass No Yes Constitutional syndrome No Yes 3 months > 3 months p 0.88 60 (48.0%) 253 (47.3%) 245 (44.5%) 68 (61.8%) 174 (49.4%) 139 (45.1%) 300 (49.0%) 13 (27.1%) 308 (47.5%) 5 (45.5%) 311 (47.4%) 2 (50.0%) 292 (47.6%) 21 (45.7%) 65 (52.0%) 282 (52.7%) 0.001 305 (55.5%) 42 (38.2%) 0.27 178 (50.6%) 169 (54.9%) 0.003 312 (51.0%) 35 (72.9%) 0.89 341 (52.5%) 6 (54.5%) 0.92 345 (52.6%) 2 (50.0%) 0.80 322 (52.4%) 25 (54.3%) 2.59 (1.344.99) 0.49 (0.330.76) OR (95% CI)

OR: odds ratio; CI: confidence interval.

As expected, the TNM stage was significantly associated with survival (p < 0.0001) (Fig. 2). Discussion The difference in survival between symptomatic and asymptomatic patients has been recognized. Some have suggested that an early diagnosis in asymptomatic patients with colorectal cancer may be an important advantage [15, 16], but it is difficult to be certain of the chronology of the symptoms associated with colorectal neoplasms. Even in prospective studies [6], there is disagreement between the dates of the clinical history and the information the patient gives during a structured interview specially designed for symptoms and dates. It is important to remember that lapses in

718

World J. Surg. Vol. 28, No. 7, July 2004

Table 3. Relation between TNM stage and the most frequently observed symptoms. TNM stage Parameter Rectal bleeding No Yes Change in bowel habit No Yes Abdominal pain No Yes Anemic syndrome No Yes 1 27 (7.7%) 72 (23.4%) 84 (15.7%) 15 (12.0%) 93 (16.9%) 6 (5.5%) 98 (16.6%) 1 (2.1%) 2 143 (40.6%) 123 (39.9%) 216 (40.4%) 50 (40.0%) 223 (40.5%) 43 (39.1%) 241 (39.4%) 25 (52.1%) 3 92 (26.1%) 70 (22.7%) 132 (24.7%) 30 (24.0%) 128 (23.3%) 34 (30.9%) 150 (24.5%) 12 (25.0%) 4 90 (25.6%) 43 (14.0%) 0.56 103 (19.3%) 30 (24.0%) 0.009 106 (19.3%) 27 (24.5%) 0.06 123 (20.1%) 10 (20.8%) p < 0.001

Table 4. Survival distributions for patients with symptoms duration < 3 months versus > 3 months (p = 0.90). Symptom duration 3 months > 3 months Overall
a

Mean survival time No. 313 347 660 Days 91.32 91.68 91.80 SE (95% CI) 5.58 (80.39102.26) 5.42 (81.06102.30) 3.92 (84.1299.48)

Median survival time Days 70.51 64.56 66.30 SE (95% CI) 14.32 (42.4598.56) 8.25 (48.3980.73) 6.55 (53.4679.14)

No. of events 137 154 291

Censoreda 176 (56.23%) 193 (55.62%) 369 (55.91%)

Censored: number of patients who were still alive at the end of the last interval of the follow-up.

memory occur more frequently in elderly people [9, 17]. Thus it is common practice to use clinical histories for studying a large number of patients with prolonged symptom evolution, and it has not created a significant problem in terms of validating the results. Usually, it is accepted that the time between the appearance of symptoms and diagnosis ranges between 3 and 6 months [6, 9, 18, 19]. Taking more than 6 months to arrive at a diagnosis is excessive and may have legal implications [20]. However, there have been reports in the literature of patients surviving who had had symptoms lasting more than 12 months [8]. A cutoff period of 3 months was considered the basis for other studies [6, 9]that had considered this length of time as diagnostic delay. Three months was the estimated median duration of symptoms in our patients. Previously, others have included patients with symptom durations of less than 12 weeks, which was an interval more favorable to a better prognosis [3]. Patients with an obstruction or perforation have had less favorable prognoses in all the studies reported, and for this reason they were not included in our study [21]. On the other hand, it has been noted that patients with a symptom duration of 3 months were less likely to have TNM stage I tumors [6, 9, 22]. Furthermore, stage IV tumors with a symptom duration of < 3 months have been associated with a poor prognosis [5]. Nevertheless, statistical analyses in our series found no significant differences between symptom duration of < 3 months versus 3 months and the TNM stage of the disease. Moreover, the two groups, homogeneous in terms of sex, age, and location of neoplasm, had no differences in long-term survival either. In fact, many authors have suggested that symptom duration is not related to survival time [23, 24]. Some even believe that prolonged duration of symptoms is associated with better survival [5, 25]. Other studies have suggested that an early tumor stage is found in patients with short symptom duration, whereas others have

found that it is the patients with intestinal obstruction and stage IV disease who have a shorter symptom duration when a singlevariable study is performed, although it does not reflect better survival rates. Poorly differentiated tumors have also been related to a poor prognosis. Data on tumor differentiation were not considered in this study because of the difficulty reported by many pathologists in distinguishing clearly and accurately between the categories of well and poorly differentiated tumors. In fact, the moderately differentiated tumor was the category most often recorded. There is a general consensus that if all the partial factors are integrated into a multivariable study only the pathologic stage is significant as a prognostic factor and determinant of survival [26]. Delays in the diagnosis of cancer are due to various factors. The factors that contribute to patients delaying seeking medical assistance [17, 27] include the type of tumor (e.g., patients with breast cancer, which is more evident, present earlier than those with rectal cancer), social class (patients in higher social classes seek help earlier because they have better information), anxiety, and pain (which makes it more likely to seek help, although generally pain is a late symptom). Incomplete examination or initial referral to a nonsurgical specialty clinic also seems to contribute to delayed diagnosis [28]. Harris and Simson reported that incomplete imaging of the colon in patients with sinister presenting symptoms was the most commonly identified factor in delayed diagnosis [29]. Likewise, the presence of abdominal symptoms due to normal physiologic processes, similar to those associated with cancer, is relatively frequent and is known. Difficulty with bowel movements, abdominal distension, and pain are found during the weeks before consultation by 10% to 20% of patients, which explains the delay in ordering colonoscopy. The presence of rectal bleeding does not necessarily indicate the presence of carcinoma. In one study, it was found in 11.8%, during the previous 3 months in healthy adults seeking help [30]. Rectal

Gonzalez-Hermoso et al.: Colorectal Cancer: Early Diagnosis and Prognosis

719

cured or, more precisely, because of the fear of having cancer. An early diagnosis during the first weeks of the symptoms should uncover a stage I or II cancer in the individual, although this was not the case in our patient group as a whole [30]. Conclusions Our observations in a large number of patients have reinforced the findings of several studies [6, 31] suggesting that when symptoms are present the prognosis of the disease has already been determined and it is not going to be changed in a significant way by the treatment used. We concluded that the duration of intestinal symptoms is not related to the stage or prognosis of the tumor. We emphasize that colorectal cancer must be diagnosed early, during the asymptomatic period, which can be achieved by means of extensive screening programs in the general population.
Re sume . Certains patients porteurs de cancer colorectal continuent a ` se pre senter a ` un stade tardif. Un retard dans le diagnostic est souvent rapporte comme cause, mais celle-ci na jamais ete valide e. Le but de cette etude a ete de de terminer si le retard du diagnostic influenc ait la survie a ` long terme et si les sympto mes les plus fre quents etaient en rapport avec le stade ou retentissait sur le moment ou ` le diagnostic a ete etabli. Ont ete analyse es re trospectivement, les donne es provenant de 660 patients traite s chirurgicalement pour cancer colorectal non complique , vus dans notre institution entre 1985 et 2000. Les variables suivantes ont ete enregistre es: a ge, sexe, sympto mes initiaux, localisation du cancer, stade TNM et survie. Les patients ont ete classe s en deux groupes selon la dure e des sympto mes: moins de trois mois vs. trois mois ou plus. Une analyse statistique comparative entre les deux groupes a ete re alise e en ce qui concernait les sympto mes initiaux, le stade TNM et la dure e de la survie. Les sympto mes initiaux les plus fre quemment rapporte s ont ete compare s au stade TNM. Les deux groupes etaient similaires eu egard la distribution de la ge, du sexe, de la location du cancer, du type de chirurgie re alise e et du stade TNM. Nous avons trouve que la dure e des sympto mes etait raccourcie en cas de douleurs abdominales (p = 0.002) (OR = 0.53; IC 95%: 0.350.80) et rallonge e en cas dane mie (p = 0.006) (OR = 2.4; IC 95%: 1.274.56). Ainsi, le stade du cancer etait corre le au saignement rectal (p < 0.001) et a ` la douleur abdominale (p = 0.008). Le test de Log-Rank a indique que la dure e des sympto mes ne tait pas en rapport avec la survie a ` long terme (p = 0.90). Nous concluons que la dure e des sympto mes du cancer colorectal ninfluence ni le stade ni le pronostic de ces tumeurs. Fig. 1. Survival curves comparing the duration of symptoms ( 3 months vs. > 3 months). Log-rank test showed no statistically significant differences between the two groups (p = 0.96). Fig. 2. Survival curves according to TNM stage (p < 0.001). Resumen. Los pacientes con ca ncer colorrectal continu an diagnostica ndose en estadios avanzados de la enfermedad. El retraso diagno stico se ha considerado un factor contribuyente, pero rara vez ha sido cuestionado. El objetivo de este estudio fue establecer si el retraso diagno stico se relacionaba con la supervivencia a largo plazo y si los sntomas ma s frecuentes se relacionaban con el estadio en el momento del diagno stico del ca ncer colorrectal. Se estudiaron retrospectivamente 660 pacientes tratados quiru rgicamente en nuestra institucio n por ca ncer colorrectal entre los an os 1995 y 2000. Se recogieron las variables edad, sexo, duracio n de los sntomas, localizacio n de la neoplasia, tipo de ciruga praticada, estado TNM y tiempo de supervivencia. Los pacientes fueron clasificados en dos grupos de acuerdo con la duracio n de los sntomas: menos de 3 meses vs ma s o igual a 3 meses. Los dos grupos se analizaron para ver si existan diferencias en cuanto al sntoma inicial, el estadio TNM y el tiempo de supervivencia. Tambie n se comparo el sntoma inicial ma s frecuente con el estadio TNM. Los dos grupos fueron similares en cuando a edad, sexo, localizacio n de la neoplasia, tipo de ciruga practicada y estadio TNM. La duracio n de los sntomas fue significativamente ma s corta en los pacientes con dolor abdominal (p = 0.002) (OR = 0.53; 95%IC: 0.350.80) y significativamente ma s alargada en los pacientes con anemia (p = 0.006) (OR = 2.4; 95%IC: 1.274.56). El estadio de la neoplasia se relaciono con la rectorragia (p < 0.001) y el dolor abdominal (p = 0.008). El test Log-Rank demostro que la duracio n de los sntomas no se relacionaba con la supervivencia a largo plazo (p = 0.90). Conclumos que la duracio n de los sntomas no se relaciona ni con el estadio ni con el prono stico del tumor a largo plazo.

bleeding is associated more often with rectal cancer (50%) than with colon cancer (14%) [11, 15]. In such cases rectoscopy or rigid sigmoidoscopy performed by a generalist probably could shorten the time to diagnosis and treatment. Some suggest that the presence of hemorrhoids implies a delay in diagnosis, especially if the patient undergoes sclerotherapy [29]. Routine colonoscopy could facilitate an early diagnosis. Other reasons for delay may be attributable to an overburdened public health care system and improper interpretation of the symptoms. All these factors explain the difficulties encountered when trying to shorten the time to diagnosis in patients with symptomatic illness. Information campaigns have not shortened the time between seeking help and the presence of symptoms indicative of colorectal cancer, perhaps because it is not advertised that the disease can be

720

World J. Surg. Vol. 28, No. 7, July 2004

References
1. Aldridge MC, Phillips RKS, Hittinger R, et al. Influence of tumor site on presentation, management and subsequent outcome in large bowel cancer. Br. J. Surg. 1986;73:663670 2. Khubchandani M. Relationship of symptom duration and survival in patients with carcinoma of the colon and rectum. Dis. Colon Rectum 1985;28:585587 3. Maguire A, Porta M, Malats N, et al. Cancer survival and the duration of symptoms: an analysis of possible forms of the risk function. ISDS II Project Investigators. Eur. J. Cancer 1994;30A:785792 4. Polissar L, Sim D, Francis A. Survival of colorectal cancer patients in relation to duration of symptoms and other prognostic factors. Dis. Colon Rectum 1981;24:364369 5. Mulcahy HE, ODonoghue DP. Duration of colorectal cancer symptoms and survival: the effects of confounding clinical and pathological variables. Eur. J. Cancer 1997;33:14611467 6. Roncoroni L, Pietra N, Violi V, et al. Delay in the diagnosis and outcome of colorectal cancer: a prospective study. Eur. J. Surg. Oncol. 1999;25:173178 7. Winawer SJ, Zauber AG, OBrien MJ, et al. Randomized comparison of surveillance intervals after colonoscopic removal of newly diagnosed adenomatous polyps: the National Polyp Study Workgroup. N. Engl. J. Med. 1993;328:901906 8. McDermott FT, Hughes ESR, Pihl E, et al. Prognosis in relation to symptom duration in colon cancer. Br. J. Surg. 1981;68:846849 9. Young CJ, Sweeney JL, Hunter A. Implications of delayed diagnosis in colorectal cancer. Aust. N. Z. Surg. 2000;70:635638 10. Pahlman L, Glimelius B, Enblad P. Clinical characteristics and their relation to surgical curability in adenocarcinoma of the rectum and rectosigmoid. Acta Chir. Scand. 1985;151:685693 11. Graffner H, Olsson S. Patients and doctors delay in carcinoma of the colon and rectum. J. Surg. Oncol. 1986;31:188190 12. Goodman D, Irvin TT. Delay in the diagnosis and prognosis of carcinoma of the right colon. Br. J. Surg. 1993;80:13271329 13. Majumdar SR, Fletcher RH, Evans AT. How does colorectal cancer present? Symptoms, duration, and clues to location. Am. J. Gastroenterol. 1999;94:30393045 14. Green FL, Page DL, Fleming ID, et al, editors. AJCC (American Joint Committee on Cancer) Cancer Staging Manual, 6th edition, New York, Springer-Verlag, 2002 15. Arbman G, Nilsson E, Storgren-Fordell V, et al. A short diagnostic 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.

delay is more important for rectal cancer than for colonic cancer. Eur. J. Surg. 1996;162:899904 Shida H, Ban K, Matsumoto M, et al. Asymptomatic colorectal cancer detected by screening. Dis. Colon Rectum 1996;39:1130115 Carter S, Winslet M. Delay in the presentation of colorectal carcinoma: a review of causation. Int. J. Colorectal Dis. 1988;13:2731 Pescatori M, Maria G, Beltrani B, et al. Site, emergency, and duration of symptoms in the prognosis of colorectal cancer. Dis. Colon Rectum 1982;25:3340 Robinson E, Hohilever J, Zidan J, et al. Colorectal cancer: incidence, delay in diagnosis and stage of disease. Eur. J. Cancer Clin. Oncol. 1986;22:157161 Kern KA. Medicolegal analysis of the delayed diagnosis of cancer in 338 cases in the United States. Arch. Surg. 1994;129:397404 Carraro P, Segala M, Cesana BM, et al. Obstructing colonic cancer: failure and survival patterns over a ten-year follow-up after one-stage curative surgery. Dis. Colon Rectum 2001;44:243250 Kemppainen M, Raiha I, Rajah T, et al. Delay in diagnosis of colorectal cancer in elderly patients. Age Ageing 1993;22:260264 Holliday HW, Hardcastle JD. Delay in diagnosis and treatment of symptomatic colorectal cancer. Lancet 1979;1:309311 Chapuis PH, Dent OF, Fisher R, et al. A multivariate analysis of clinical and pathological variables in prognosis after resection of large bowel cancer. Br. J. Surg. 1985;72:698702 McDermott F, Hughes E, Pihl E, et al. Symptom duration and survival prospects in carcinoma of the rectum. Surg. Gynecol. Obstet. 1981;153: 321326 Pietra N, Sarli L, Thenasseril BJ, et al. Risk factors of local recurrence of colorectal cancer: a multivariate study. Hepatogastroenterology. 1998;45:15731578 Funch DP. Diagnostic delay in symptomatic colorectal cancer. Cancer 1985;56:21202124 Potter MA, Wilson RG. Diagnostic delay in colorectal cancer. J. R. Coll. Surg. Edinb. 1999;44:313316 Harris GJ, Simson JN. Causes of late diagnosis in cases of colorectal cancer seen in a district general hospital over a 2-year period. Ann. R. Coll. Surg. Engl. 1998;80:246248 Dent OF, Goulston KJ, Tennant CC, et al. Rectal bleeding: patient delay in presentation. Dis. Colon Rectum 1990;33:851857 Kiran PR, Glass RE. Duration of symptoms and spread of colorectal cancer: a short history does not mean early disease. Ann. R. Coll. Surg. Engl. 2002;84:381385