| |
Screening colonoscopy for the detection of neoplastic lesions in asymptomatic HIV-infected subjects: more colon lesions in HIV+
|
| |
| |
Gut Sept 2009;58:1129-1134
E J Bini1,2, B Green2, M A Poles1,2
1 Division of Gastroenterology, VA New York Harbor Healthcare System, New York, USA
2 Department of Medicine, New York University School of Medicine, New York, USA
Dr E J Bini, Division of Gastroenterology (111D), VA New York Harbor Healthcare System, 423 East 23rd Street, New York, NY 10010, USA;
Edmund.Bini@med.va.gov
"The prevalence of neoplastic lesions detected by colonoscopy was significantly higher in HIV-infected patients than in control subjects (62.5% vs 41.2%, p<0.001), a difference of 21.3% (95% CI, 11.2% to 31.5%). The proportion of HIV positive subjects with neoplastic lesions was significantly higher than in the control subjects even after excluding those individuals who had a colonoscopy >10 years ago or a flexible sigmoidoscopy >5 years ago (64.7% vs 42.0%, p<0.001)."
"Although we and others have shown that colorectal cancer develops at a younger age and is more aggressive in patients with HIV as compared with uninfected controls, the reasons for these findings are unclear. One possible explanation for the more aggressive nature of colorectal cancer in subjects with HIV is that the altered host immune response in this population may permit earlier development of, and a more aggressive course, of colorectal cancer. This hypothesis is supported by numerous studies demonstrating that AIDS-defining as well as non-AIDS-defining malignancies are more common in subjects with HIV than in the general population"
"in conclusion, our findings suggest that screening colonoscopy should be offered to HIV-infected subjects. The high prevalence of colonic neoplasms in patients with HIV has important clinical and policy implications for the future delivery of preventive medicine to this population, and highlights the need to increase awareness of colorectal cancer screening among subjects with HIV and their healthcare providers. Future studies of HIV-infected subjects are needed to determine the appropriate age to begin screening for colorectal cancer, the optimal frequency of screening, the identification of HIV-infected subjects who are most likely to benefit from screening, and the cost-effectiveness of screening for colorectal cancer in this population."
"HIV-infected patients had a significantly increased odds of having a neoplastic lesion (OR = 2.38; 95% CI, 1.56 to 3.63), and this association remained highly significant after adjustment for age, sex, and race/ethnicity (OR = 2.34; 95% CI, 1.49 to 3.68) and after adjustment for all baseline characteristics listed in table 1 (OR = 3.00; 95% CI, 1.83 to 4.93)."
"Compared with control subjects, HIV-infected patients were significantly less likely to have hyperplastic polyps and were more likely to have adenomas 6-9 mm in diameter. Advanced neoplastic lesions and adenocarcinomas of the colon were more common in HIV-infected subjects, although these differences did not reach statistical significance. More HIV-infected subjects than control subjects had two or more adenomas detected (41.2% vs 30.9%)"
"Among the 11 patients who were diagnosed with adenocarcinoma, HIV-infected subjects were significantly younger than those without HIV (52.4 (SD 1.3) vs 60.3 (SD 4.0) years, p = 0.002), a difference of 7.9 (95% CI, 3.6 to 12.2) years. Late-stage adenocarcinoma of the colon (stage III or IV) was more common in HIV-infected subjects (three of five (60.0%)) than in uninfected control subjects (one of six (16.7%)), although this difference was not statistically significant (p = 0.24)."
ABSTRACT
Background: Although non-AIDS defining malignancies are rapidly increasing as HIV-infected subjects live longer, little is know about the results of screening for colonic neoplasms (adenomatous polyps and adenocarcinomas) in this population.
Methods: We conducted a screening colonoscopy study to determine the prevalence of colonic neoplasms in 136 asymptomatic HIV-infected subjects >=50 years of age and 272 asymptomatic uninfected control subjects matched for age, sex, and family history of colorectal cancer. Advanced neoplasms were defined as adenomas >=10 mm or any adenoma, regardless of size, with villous histology, high-grade dysplasia, or adenocarcinoma.
Results: The prevalence of neoplastic lesions was significantly higher in HIV-infected subjects than in control subjects (62.5% vs 41.2%, p<0.001), and remained highly significant after adjustment for potential confounding variables (odds ratio = 3.00; 95% confidence interval, 1.83 to 4.93). Among patients with colorectal adenocarcinoma, HIV-infected subjects were significantly younger (52.4 (SD 1.3) vs 60.3 (SD 4.0) years, p = 0.002) and were more likely to have advanced cancers (stage III or IV) than control subjects (60.0% vs 16.7%, p = 0.24). Of HIV-infected subjects with advanced neoplasms proximal to the splenic flexure, distal neoplastic lesions were absent in 88.9% of individuals and these would have been missed by flexible sigmoidoscopy.
Conclusions: HIV-infected subjects have a higher prevalence of colonic neoplasms, and adenocarcinomas develop at a younger age and are more advanced than in uninfected subjects. Our findings suggest that screening colonoscopy should be offered to HIV-infected subjects, but the age of initiation and the optimal frequency of screening require further study.
Highly active antiretroviral therapy (HAART) has had a substantial impact on the natural history of human immunodeficiency virus (HIV) infection and acquired immunodeficiency syndrome (AIDS) and has led to a marked decrease in mortality.1-6 As a result, HIV infection is now a chronic disease and the life expectancy of patients with HIV continues to increase.7-11
In addition to improving the life expectancy of patients with HIV, HAART has also been associated with a significant decrease in the incidence of AIDS-defining malignancies such as Kaposi's sarcoma, cervical cancer, and non-Hodgkin's lymphoma.12-14 In contrast, non-AIDS defining malignancies are on the rise and have become an increasingly important cause of death in patients with HIV.13 15-19 Numerous studies have demonstrated that the risk of developing non-AIDS defining malignancies is higher in HIV-infected patients compared with the general population.20-30
As HIV-infected patients age, it is anticipated that there will be a further increase in the incidence of non-AIDS defining malignancies such as cancers of the colon, prostate and breast, and screening of this population for these malignancies may lead to improved survival.31 32 Although screening for colorectal cancer is currently recommended by the U.S. Preventive Services Task Force and other professional organisations for all average-risk individuals starting at 50 years of age,33-37 colonoscopy has not been evaluated as a screening test in HIV-infected persons. Therefore, we conducted a prospective screening colonoscopy study to evaluate the prevalence of colorectal neoplasms in HIV-infected individuals. Since chronic immunosuppression is thought to increase the risk of developing numerous non-AIDS defining malignancies, we hypothesised that HIV-infected individuals would be more likely than non-infected individuals to have colorectal neoplasms detected by screening colonoscopy.
METHODS
Study population
We prospectively identified all patients referred for screening colonoscopy at the Veterans Affairs (VA) New York Harbor Healthcare System in New York City from April 2002 through October 2004. Patients were eligible if they were >=50 years of age and were asymptomatic. Asymptomatic was defined as the absence of lower gastrointestinal symptoms, including abdominal pain, constipation, diarrhoea, change in stool calibre or frequency, unexplained weight loss, rectal bleeding, or other alarm symptoms suggestive of an underlying malignancy.
Patients were excluded if they had previously undergone acceptable methods of colorectal cancer screening, including flexible sigmoidoscopy or barium enema within the last 5 years or a colonoscopy within the last 10 years. Subjects were also excluded if they had a positive faecal occult blood test.
For each HIV-infected patient who was eligible for this study, we identified the next two consecutive asymptomatic HIV negative patients who underwent a screening colonoscopy. The HIV negative control subjects were matched with the HIV-infected subjects for age (within 1 year), sex, and family history of colorectal cancer in a first-degree relative.
Study design
Demographic and clinical information were obtained from all patients prior to colonoscopy. Information collected included age, sex, self-reported race/ethnicity, use of alcohol or tobacco, use of non-steroidal anti-inflammatory drugs (NSAIDs) or aspirin, prior colonoscopy >10 years ago, and flexible sigmoidoscopy >5 years ago. For HIV-infected patients, we also collected data on the duration of infection and the antiretroviral medications taken. In addition, we reviewed the electronic medical records of all HIV-infected subjects to confirm their HIV status and to obtain the results of plasma HIV RNA testing and CD4 lymphocyte counts closest to the time of colonoscopy.
All patients received an oral sodium phosphate solution or a polyethylene glycol-based electrolyte solution for bowel preparation. After administration of intravenous meperidine and midazolam for conscious sedation, the colon was evaluated to the caecum and the location and size of each polyp was recorded. The endoscopists who performed the colonoscopies were blinded to the aims of this study and were instructed to remove all visible polyps. The size of small polyps was estimated using an open-biopsy forceps, which is 7 mm in diameter. The size of large polyps removed by snare polypectomy was confirmed during histologial evaluation. All retrieved polyps and biopsy specimens were sent to our local pathology laboratory for histological evaluation.
Study outcomes
The primary outcome was the prevalence of neoplastic lesions (adenomatous polyps and adenocarcinomas) detected by colonoscopy among HIV positive subjects as compared with control subjects. Assuming a prevalence of neoplastic lesions of 50% in the HIV-infected patients and 35% in the control subjects, 127 HIV-infected patients and 254 control subjects would be needed to detect a 15% difference at an alpha of 0.05 and a power of 80%. The estimated prevalence of neoplastic lesions in the control subjects was based upon published data from the VA Cooperative Study Group 380 screening colonoscopy study.38
The secondary outcomes of this study included the proportion of HIV-infected patients and control subjects with advanced neoplasia, the size of the adenomas detected (<=5 mm, 6-9 mm, and >=10 mm in diameter), the stage of colorectal cancer, and the distribution of the colonic neoplasms (proximal vs distal colon). Advanced neoplasia was defined as adenomas >=10 mm in diameter or any adenoma, regardless of size, with villous histology, high-grade dysplasia, or cancer.38 Colorectal cancers were staged using the TNM classification system proposed by the American Joint Committee on Cancer.39 The distal colon was defined as the rectum, sigmoid, and descending colon up to, but not including, the splenic flexure, thus within the reach of a standard flexible sigmoidoscope. The proximal colon was defined as the segment of colon proximal to, and including, the splenic flexure. In addition, we evaluated demographic and clinical factors associated with neoplastic lesions in the subjects infected with HIV.
Statistical analysis
Continuous variables are expressed as mean with the SD for those variables that were normally distributed, and medians and interquartile range (IQR; 25th to 75th percentiles) for those with a non-normal distribution. Continuous variables were compared using the unpaired two-tailed t test or the Mann-Whitney U test. Categorical variables are expressed as proportions and were compared using the {chi}2 test or Fisher's exact test.
Multivariable logistic regression analysis was used to assess the effect of HIV infection on the odds of having a neoplastic lesion after adjustment for potential confounding variables. The potential confounding variables that were included in the multivariable logistic regression model included age, sex, race/ethnicity, current alcohol use, current smoking, use of NSAIDs and aspirin, family history of colorectal cancer in a first-degree relative, colonoscopy >10 years ago, and flexible sigmoidoscopy >5 years ago. The strength of the association between HIV infection and the presence of colonic neoplasia is expressed as odds ratios (OR) with 95% confidence intervals (CIs). In addition, we utilised simple logistic regression analysis to determine factors associated with having a colonic neoplasm in the HIV-infected patients. All variables that were significant (p<0.05) in the simple logistic regression analysis were included in the multivariable logistic regression analysis. The assumptions of multivariable logistic regression analysis were checked and satisfied for all models. Statistical analysis was performed using SPSS software version 15.0 for Windows and a two-tailed p value of <0.05 was considered statistically significant.
RESULTS
Patient demographic and clinical characteristics
A total of 136 HIV-infected patients and 272 uninfected control subjects were enrolled. The baseline demographic and clinical characteristics of the study subjects are shown in table 1. Compared with control subjects, HIV-infected patients were significantly more likely to be racial/ethnic minorities and were more likely to smoke, whereas they were less likely to currently be using alcohol or taking aspirin.
Among the HIV-infected patients, the median duration of infection was 11.0 years (IQR, 7.0-14.0 years), and 85 (62.5%) were diagnosed with HIV >=10 years ago. The median CD4 lymphocyte count was 346.0 cells/mm3 (IQR, 236.8-707.8 cells/mm3), with 69 (50.7%) having a CD4 count below 350 cells/mm3 and 23 (16.9%) having a CD4 count below 200 cells/mm3. A total of 124 of the 136 HIV-infected subjects (91.2%) were taking HAART, and plasma HIV RNA was undetectable in 100 subjects (73.5%).
Prevalence of neoplastic lesions
Colonoscopy was complete to the caecum in 131 (96.3%) of the 136 HIV-infected patients and 266 (97.8%) of the 272 control subjects (p = 0.52). The prevalence of neoplastic lesions detected by colonoscopy was significantly higher in HIV-infected patients than in control subjects (62.5% vs 41.2%, p<0.001), a difference of 21.3% (95% CI, 11.2% to 31.5%). The proportion of HIV positive subjects with neoplastic lesions was significantly higher than in the control subjects even after excluding those individuals who had a colonoscopy >10 years ago or a flexible sigmoidoscopy >5 years ago (64.7% vs 42.0%, p<0.001).
Compared with control subjects, HIV-infected patients had a significantly increased odds of having a neoplastic lesion (OR = 2.38; 95% CI, 1.56 to 3.63), and this association remained highly significant after adjustment for age, sex, and race/ethnicity (OR = 2.34; 95% CI, 1.49 to 3.68) and after adjustment for all baseline characteristics listed in table 1 (OR = 3.00; 95% CI, 1.83 to 4.93). To address the possibility of residual confounding by using dichotomous variables for aspirin use, NSAID use, smoking, and alcohol use, we repeated the fully adjusted model using the number of doses of aspirin and NSAIDs consumed per week, packs of cigarettes smoked per day, and number of drinks of alcohol consumed daily and the association between HIV infection and neoplastic lesions did not change substantially (OR = 2.84; 95% CI, 1.74 to 4.62).
Types of neoplastic lesions detected by colonoscopy
The types of lesions identified by colonoscopy are shown in fig 1. Compared with control subjects, HIV-infected patients were significantly less likely to have hyperplastic polyps and were more likely to have adenomas 6-9 mm in diameter. Advanced neoplastic lesions and adenocarcinomas of the colon were more common in HIV-infected subjects, although these differences did not reach statistical significance. More HIV-infected subjects than control subjects had two or more adenomas detected (41.2% vs 30.9%, p = 0.04).
Figure 1 Prevalence of neoplastic lesions in the HIV-infected subjects and control subjects. Note: These categories are not mutually exclusive since advanced neoplasia includes all patients with adenomas >=10 mm, those with adenomas of any size with villous histology or high-grade dysplasia, and individuals with adenocarcinoma.
Distribution of the colonic neoplasms
HIV-infected subjects were significantly more likely than control subjects to have one or more neoplastic lesions in the proximal colon (37.5% vs 24.3%, p = 0.005) and in the distal colon (56.6% vs 30.9%, p<0.001). To determine the likelihood that proximal neoplastic lesions would have been missed by only examining the distal colon, we evaluated the absence of neoplastic lesions distal to the splenic flexure (within the reach of the flexible sigmoidoscope). Among the 117 subjects with proximal neoplastic lesions, 15.7% of HIV-infected subjects and 42.4% of control subjects did not have any neoplastic lesions in the distal colon (p = 0.002). Of the 27 subjects with advanced neoplasms in the proximal colon, 88.9% of HIV-infected subjects and 33.3% of uninfected control subjects did not have any neoplastic lesions in the distal colon (p = 0.01).
Factors associated with neoplastic lesions among subjects with HIV
To evaluate factors associated with colorectal neoplastic lesions among the 136 HIV-infected subjects, we determined the proportion of subjects who had neoplastic lesions detected by colonoscopy according to select demographic and clinical characteristics (table 2). Using simple logistic regression analysis we found that a family history of colorectal cancer in a first-degree relative and current use of HAART were the only variables that was significantly associated with having a neoplastic lesion detected by colonoscopy. In contrast, there was no association between neoplastic lesions and the duration of HIV infection, HIV viral load, or CD4 lymphocyte count. Multivariable logistic regression analysis identified that the odds of having a neoplastic lesion was higher in those with a family history of colorectal cancer (OR = 3.77; 95% CI, 1.20 to 11.86) and lower in those taking HAART (OR = 0.13; 95% CI, 0.02 to 1.02).
DISCUSSION
In the present study, we demonstrated that neoplastic lesions detected by colonoscopy were significantly more common in HIV-infected subjects than in age-, sex- and family history-matched control subjects without HIV. This clinically important finding has substantial implications for the estimated 1 million persons living with HIV in the United States,40 especially since at least 20% of all HIV-infected subjects are >=50 years of age.41-43
Although the life expectancy of HIV-infected patients has increased dramatically, little is known about colorectal cancer screening in this population. In a study of flexible sigmoidoscopy in 165 HIV-infected subjects and 2217 subjects without HIV, we previously demonstrated that the prevalence of neoplastic lesions (25.5% vs 13.1%) and advanced neoplasms (7.3% vs 3.8%) of the distal colon were significantly higher in subjects with HIV.44
The present screening colonoscopy study expands upon our prior work and demonstrates that colonoscopy can detect neoplastic lesions in an even greater proportion of HIV-infected subjects (62.5%) as compared with the 25.5% of neoplasms that were detected by flexible sigmoidoscopy in our prior study.44 Furthermore, we demonstrated that 15.7% of all neoplastic lesions and 88.9% of advanced neoplastic lesions in the proximal colon of HIV-infected subjects likely would have been missed if flexible sigmoidoscopy had been performed instead of colonoscopy. These findings suggest that colonoscopy may be superior to flexible sigmoidoscopy to screen for colorectal cancer in subjects with HIV, although this requires confirmation in future studies.
In addition to demonstrating that neoplastic lesions of the colon are more common in subjects with HIV as compared with control subjects, we also found that HIV-infected persons with adenocarcinomas of the colon were younger and had more advanced disease as compared with HIV-negative control subjects. These findings are in agreement with several published case reports,45 46 case series,47 48 and a cancer registry study.49 In the study by Yeguez et al,47 four of the six HIV-negative subjects with adenocarcinoma of the colon were <=40 years of age and five of the six subjects had metastatic disease at the time of presentation. Given the findings that colorectal cancers develop at an earlier age and are more aggressive in HIV-infected subjects as compared with those without HIV, it is critical that future studies determine the appropriate age to begin colorectal cancer screening as well as the optimal frequency of screening.
Although we and others have shown that colorectal cancer develops at a younger age and is more aggressive in patients with HIV as compared with uninfected controls, the reasons for these findings are unclear. One possible explanation for the more aggressive nature of colorectal cancer in subjects with HIV is that the altered host immune response in this population may permit earlier development of, and a more aggressive course, of colorectal cancer. This hypothesis is supported by numerous studies demonstrating that AIDS-defining as well as non-AIDS-defining malignancies are more common in subjects with HIV than in the general population.20-30
However, we did not find any association between neoplastic lesions of the colon and either the duration of HIV infection, CD4 lymphocyte count, or HIV viral load. In contrast, we did find that use of HAART was associated with a significantly lower odds of colonic neoplasms. The reasons for these findings are unknown and require further investigation.
The strengths of our study include the prospective study design, our collection of detailed demographic and clinical data, the inclusion of a control group of subjects without HIV, and the performance of complete colonoscopy in nearly all subjects. In addition, we are unaware of any published studies that have evaluated the utility of screening colonoscopy in HIV-infected subjects.
However, there are several limitations to consider when interpreting our findings. First, the study was conducted at a single VA Medical Center and the majority of our patients were men. Therefore, our findings may not be generalisable to other clinical settings or to women. Second, our study population included only 136 HIV-infected patients and larger prospective studies are needed to address this clinically important issue.
Third, although HIV-infected subjects and control subjects were matched by age, sex, and family history of colon cancer, there were still significant differences in other baseline demographic and clinical characteristics between the two groups. Although the association between HIV infection and colonic neoplasms remained statistically significant even after adjustment for these baseline characteristics, it is possible that there were unmeasured differences between the two groups of subjects.
Finally, HIV infection was well controlled in the majority of our HIV-infected subjects, with 83.1% having a CD4 lymphocyte count of at least 200 cells/mm3 and 73.5% having undetectable HIV RNA. Therefore, our findings may not be generalisable to subjects with poorly controlled HIV infection, a population who is likely to have a much shorter life expectancy, but for whom colorectal cancer screening may not be appropriate.
In conclusion, our findings suggest that screening colonoscopy should be offered to HIV-infected subjects. The high prevalence of colonic neoplasms in patients with HIV has important clinical and policy implications for the future delivery of preventive medicine to this population, and highlights the need to increase awareness of colorectal cancer screening among subjects with HIV and their healthcare providers. Future studies of HIV-infected subjects are needed to determine the appropriate age to begin screening for colorectal cancer, the optimal frequency of screening, the identification of HIV-infected subjects who are most likely to benefit from screening, and the cost-effectiveness of screening for colorectal cancer in this population.
|
|
| |
| |
|
|
|
