These two authors contributed equally to this work.
Department of Internal Medicine I, Bonn University Hospital, Bonn, Germany
Correspondence: Dr Christoph Boesecke, Department of Internal Medicine I, Bonn University Hospital, Sigmund-Freud-Str. 25, 53127 Bonn, Germany. Tel: +49 228 287 16558; fax: +49 228 287 15034; e-mail: email@example.com
Because of the improved life expectancy provided by successful antiretroviral combination therapy, preventive health measures in HIV-infected patients have assumed increasing importance. To date, no data exist on rates of mucosal abnormalities detected by screening colonoscopy in > 50-year-old HIV-infected patients in Germany. The aim of this study was to obtain such data.
A screening colonoscopy was offered to 159 HIV-infected patients (age > 50 years) who presented for HIV standard of care visits at the infectious diseases out-patient clinic at the university hospital in Bonn over a 1-year period from February 2010. Pearson's χ2 test, Fisher's exact test and the Mann−Whitney U-test were used for statistical analysis.
Fifty-one patients (32.1%) had undergone a screening colonoscopy in the past 10 years, and 45 patients (28.3%) were eventually screened in the observation period. The median age of the 96 screened patients (86% male and 14% female) was 58 years [interquartile range (IQR) 54–64 years]. Overall, endoscopic abnormalities were found in 61% of patients. Histological examination showed tubular adenomas in 21.9% of patients, tubulovillous adenomas in 3.1% and serrated adenomas in 1%. Hyperplastic polyps were found in 15.6% of patients, a nonspecific colitis in 16.7% and diverticulosis in 12.5%. In four cases there was even an early-stage carcinoma (two anal, one rectal and one colon cancer). In univariate analysis, no significant differences with regard to immune status, highly active antiretroviral therapy, family history, personal risk factors or comedication were found between patients with dysplastic and normal mucosas.
The high acceptance rate of screening colonoscopy and the in comparison with the HIV-negative population comparably higher rate of abnormalities in this cohort of HIV-infected patients justify enhanced implementation of screening colonoscopy in clinical practice.
Colorectal cancer (CRC) is the second most common type of cancer and the second most common cause of death from cancer in men and women in Germany. The epidemiological cancer registry in Germany recorded 68 740 new cases and 27 225 deaths attributable to CRC in 2006 . Since October 2002, screening colonoscopy has been offered to persons aged 55 years and older as part of CRC screening in Germany .
As a result of effective antiretroviral medication, HIV-infected patients now have a longer life expectancy [3-5]. The introduction of antiretroviral therapy (ART) has also led to a change in the spectrum of malignancies in HIV-infected patients. HIV-infected patients are more likely to develop non-AIDS-defining cancers such as Hodgkin lymphoma, lung cancer and adenocarcinoma of the colon [6-14]. Consequently, the number of deaths attributable to non-AIDS-defining malignancies has been increasing, while the number of AIDS-related deaths has declined [3, 15-18].
As the incidence of CRC increases rapidly above the age of 50 years, the older HIV-positive population may show an increased risk of developing CRC . There are limited data on the association of CRC and HIV infection and also a lack of screening data in the HIV-positive population. Relative to HIV-negative patients, HIV-infected patients seem to have a lower age at the time of CRC diagnosis, poorer grading, less favourable Dukes stage and a shorter survival time [19-21]. In addition, CRC screening is less frequently performed in the HIV-infected population . As screening reduces CRC mortality by detection and removal of slow-growing precursor lesions, CRC screening should be offered to HIV-infected subjects in a timely manner.
To date, there are no sustainable data on CRC screening colonoscopy in the HIV-infected population in Germany. Our primary objective was therefore to examine the acceptance of screening colonoscopy, the prevalence of colonic neoplasms and abnormalities, and the influence of antiretroviral therapy, protective medication and other factors associated with the development of CRC on mucosal dysplasia in a large cohort of elderly HIV-infected patients.
A total of 159 of 231 HIV-infected patients aged 50 years and older who presented for HIV standard of care visits at the infectious diseases out-patient clinic at the university hospital in Bonn were offered a screening colonoscopy over a period of 12 months from February 2010 to February 2011. Patients with an indication for a therapeutic colonoscopy, a known colonic disease (e.g. colitis ulcerosa) or clinical symptoms (e.g. diarrhoea) were excluded from the study. Suspicious mucosal tissue seen during colonoscopy was biopsied and a histopathological examination was performed subsequently. MOVIPREP® (Norgine, Marburg, Germany) was used for preparation.
A detailed medical history was obtained, including epidemiological parameters (age, gender, transmission risk and date of diagnosis of HIV infection), clinical parameters (body mass index, results of histological examination, CDC stage, comorbidities, coinfections, comedications and family history), antiretroviral therapy (duration and agents) and laboratory markers (CD4 count and HIV RNA).
Pearson's χ2 test, Fisher's exact test, the Mann−Whitney U-test and descriptive statistical methods [median and interquartile range (IQR)] were used for statistical analysis. The results of the statistical tests were considered significant when the two-sided P-value did not exceed 0.05. spss software for Windows (release 18.0; IBM, Armonk, NY, USA) was used for statistical analysis.
Of the 159 HIV-infected patients who were offered CRC screening colonoscopy over a 12-month period, 136 were male (85.5%) and 23 were female (14.5%). The median age was 58 years, with the youngest patient being 52 and the oldest being 84 years old. Most patients were in the age groups 50–54 (30.2%), 55–59 (28.9%), 60–64 (15.1%) and 65–69 (15.1%) years. In total, 13.8% of patients had a positive family history for CRC.
Only 24 of the 159 patients (12.4%) rejected the screening colonoscopy. Sixteen patients (10.1%) who needed time for consideration and 23 patients (14.5%) who agreed to a screening colonoscopy did not have a colonoscopy performed in the observation period.
Fifty-one patients (32.1%) had undergone a screening colonoscopy in the past 10 years and 45 patients (28.3%) were eventually screened in the observed period (Fig. 1).
A screening colonoscopy was performed in 96 patients, of whom 85.4% were male and 14.6% were female. The median age of these patients was 58 years. Most patients were in the age groups 55–59 (30.2%), 50–54 (27.1%) and 60–64 (18.8%) years (Fig. 2). For two-thirds of the patients (68.8%) it was the first colonoscopy ever, and for 26% it was the second. The quality of preparation was excellent; only three patients had soiled mucosa and had to undergo a second colonoscopy.
The main transmission risk for HIV was men who have sex with men (MSM; 40.6%), followed by heterosexual transmission (22.9%), injecting drug use (IDU; 7.3%) and transfusion (6.3%). In 22.8% of patients the transmission risk was unknown.
Most patients (95.6%) were on highly active antiretroviral therapy (HAART), and the median duration of HAART was 9 years (IQR 3–14 years). As a result, the median CD4 T-cell count was high, at 499 cells/μL (IQR 338–655 cells/μL), and 74.6% had HIV RNA below the limit of detection (40 HIV-1 RNA copies/ml).
A total of 16.5% of patients were coinfected with hepatitis C virus, 4.3% had chronic hepatitis B virus infection and 14.1% self-reported a history of genital condylomata. The prevalence of comorbidities was as follows: cardiovascular diseases were present in 38.5% of patients, diabetes mellitus in 13.5%, obesity in 10.4%, haemophilia in 4%, chronic inflammatory bowel disease in 1% and other types of comorbidity in 85.7%. Regarding medication known to be potentially protective against CRC, 19.8% of patients used acetylsalicylic acid and 15.6% statins. Other medications were used by 71.9% of patients, mostly for treatment of cardiovascular disease (Fig. 3).
Overall, in 61.7% of patients endoscopic abnormalities of any quality were found. Most patients with endoscopic abnormalities were male (87.9%), as were those without abnormalities (83.3%). The median age in both groups was 58 years. In detail, patients with abnormalities were mainly in the age groups 55–59 (29.3%), 50–54 (25.9%) and 60–64 (19%) years. It was the first screening colonoscopy for most patients in both groups. A positive family history for CRC was known in 17.2% of patients with endoscopic abnormalities compared with 11.1% of patients without abnormalities, with no statistically significant difference for occurrence of endoscopic abnormalities between patients with and without a family history of CRC (P = 0.146). The parameters gender (P = 0.522), age (P = 0.682), transmission risk (P = 0.183) and number of colonoscopies (P = 0.684) did not show statistically significant associations with endoscopic abnormalities either. Most patients with (96.3%) and without (94.3%) endoscopic abnormalities were treated with HAART. The median duration of HAART treatment was 10 years (IQR 4–14 years) in patients with endoscopic abnormalities and 7 years (IQR 3–12 years) in patients without abnormalities. Neither the duration of HAART (P = 0.056) nor HAART status (P = 0.644) showed a statistically significant association with the occurrence of endoscopic abnormalities. Also, no significant association between endoscopic abnormalities and cumulative exposure to the most commonly used antiretrovirals, such as emtricitabine (P = 0.8), didanosine (P = 0.099), lamivudine (P = 0.693), abacavir (P = 0.2), zidovudine (P = 0.143), stavudine (P = 0.835), efavirenz (P = 0.108), nevirapine (P = 0.913), darunavir (P = 0.905), atazanavir (P = 0.739), lopinavir (P = 0.474), ritonavir (P = 0.699), fosamprenavir (P = 0.5), nelfinavir (P = 1) and saquinavir (P = 0.368), was seen. Only the usage of tenofovir was borderline significant with regard to development of mucosal alterations (P = 0.057).
Overall, 61.7% of patients with endoscopic abnormalities underwent biopsy. Histological examination revealed adenomas in 25 (26.0%) of screened patients, tubular adenoma in 21 (21.9%), tubulovillous adenoma in three (3.1%) and serrated adenomas in one (1%), resulting in an overall dysplasia frequency of 26.1% in the screened population (Fig. 4). In addition, nonspecific colitis was found in 16 patients (16.7%), hyperplastic polyps in 15 (15.6%) and diverticulosis in 12 (12.5%). In four patients, cancerous lesions were detected (two anal cancers, one colon cancer and one rectal cancer) (Fig. 4).
Most patients with dysplastic mucosa were male (87.5%). The median age of the patients was 60 years (IQR 56–67 years), in contrast to the median age of 58 years (IQR 53–62 years) in patients without dysplastic mucosa (P = 0.098). In detail, patients with dysplastic mucosa were mainly in the age groups 55–59 (29.2%), 65–69 (25%), 50–54 (16.7%) and 60–64 (16.7%) years (Fig. 5). At screening, most patients were classified as being in Centers for Disease Control and Prevention (CDC) category A (38.1%), but 33.4% were in CDC category C. It was the first screening colonoscopy for most patients in both groups. A positive family history for CRC was reported by 12.5% of the patients with dysplastic mucosa compared with 16.2% of patients without dysplastic mucosa, with no statistically significant difference for occurrence of dysplastic mucosa between patients with and without a family history of CRC (P = 0.420). The parameters gender (P = 1) and age (P = 0.098) did not show statistically significant associations with the occurrence of dysplastic mucosa either. Most patients with (90.9%) and without (96.9%) dysplastic mucosa were treated with HAART. The median duration of HAART was 8 years (IQR 3–15 years) in patients with dysplastic mucosa and 9 years (IQR 4–14 years) in patients without dysplastic mucosa. Neither the duration of HAART (P = 0.823) nor HAART status (P = 0.264) showed a statistically significant association with the occurrence of dysplastic mucosa. In univariate analysis, cumulative exposure to the antiretrovirals didanosine (P = 0.013) and efavirenz (P = 0.033) was statistically significantly associated with the occurrence of dysplastic mucosa. No statistically significant association was found for the use of tenofovir (P = 0.195), lamivudine (P = 0.956), abacavir (P = 0.384), zidovudine (P = 0.908), stavudine (P = 0.735), nevirapine (P = 0.343), darunavir (P = 0.905), atazanavir (P = 0.437), lopinavir (P = 0.849), ritonavir (P = 0.530) and fosamprenavir (P = 0.1).
There were no statistically significant differences with regard to comorbidities, comedications, immune status or personal risk factors between patients with and without dysplastic mucosa.
To our knowledge, this is the first cohort study reporting on colonoscopy results as part of a CRC screening strategy in HIV-infected patients in Germany. Firstly, our study showed a high level of acceptance for screening colonoscopy in HIV-infected patients of 60.4%. For comparison, the acceptance for screening colonoscopy in the general population in Germany is about 3% annually and 27–44% in a 10-year period [22, 23], far lower than the rate seen in our cohort. Regular medical care and a close patient−physician relationship in the setting of a chronic illness such as HIV infection may lead to HIV-infected patients being well educated regarding cancer in general and CRC and screening in particular, resulting in a higher acceptance rate. Our findings are in contrast to data from US studies where lack of use of preventive measures for CRC in HIV-infected patients compared with the general population has been reported [21, 24, 25].
In addition, the adenoma detection rate during CRC screening colonoscopy in HIV-positive patients appears to be higher than that seen in the general population in Germany. Our data revealed adenomas in 26% of all screened HIV-positive patients, while published data predict adenoma detection rates of 12–20% in the general population [23, 26, 27]. Recently published studies from the USA suggest a higher incidence of colorectal adenomas in the HIV-infected population compared with the general population, which is consistent with our observations [20, 25, 28]. Bini et al. showed a significantly higher prevalence of neoplastic lesions in HIV-positive patients compared with HIV-negative control subjects (62.5% vs. 41.2%, respectively; P < 0.001) . Data from another US study, by Iqbal et al., were consistent with these results (50% vs. 23.8%, respectively; P = 0.0281) .
In addition, > 50-year-old HIV-infected patients were less likely to have ever had at least one CRC screening test compared with HIV-negative patients in the study by Reinhold et al. (55.6% vs. 77.8%, respectively; P < 0.001), especially screening colonoscopy (17.2% vs. 27.5%, respectively; P = 0.002) . This lower examination rate is confirmed by Iqbal et al., showing an underutilization of CRC screening in HIV-infected patients, with 41% of HIV-infected patients being up to date for CRC screening compared with 67% of HIV-negative patients (P = 0.0001) .
Long-term antiretroviral therapy, effective viral suppression, known protective medication for CRC and the majority of antiretroviral drugs were not found to have much influence on the occurrence of mucosal dysplasia in the patients in our cohort. These findings are consistent with those of some other studies. Berretta et al. and Wasserberg et al. were unable to establish a correlation between CD4 count, HIV viral load and stage of CRC [19, 30]. This is in contrast to recently published data showing a lower risk for CRC in HIV-infected patients on HAART under the assumption of a correlation between low immune status and non-AIDS-defining cancers [20, 29, 31-33]. In their studies, Bini et al. showed significant associations between CD4 count and distal neoplastic lesions of the colon, with the highest rates of neoplastic lesions in patients with CD4 counts < 200 cells/μL (44.7%) [20, 29]. The authors also demonstrated that the odds of having a neoplastic lesion were lower in patients taking HAART [odds ratio (OR) 0.13; 95% confidence interval (CI) 0.02 to 1.02]. Patel et al. also showed an association between antiretroviral therapy and decreased risk for CRC (relative risk 0.5; P = 0.027) . In addition, Kesselring et al. and Prosperi et al. showed a correlation between low CD4 count and increased risk of several non-AIDS-defining malignancies [31, 32].
The association between cumulative exposure to two antiretroviral drugs and the occurrence of mucosal dysplasia needs to be interpreted with caution. Firstly, our cohort study was nonrandomized and, secondly, the two drugs are given as part of a combination therapy which does not allow the influence of a single agent to be determined.
Screening colonoscopy has been offered to the general population as part of cancer screening in Germany since October 2002. It is available to men and women aged 55 years and older. However, as 16.7% of all screened patients in our cohort were younger than 55 years and the rate of dysplasia was relatively high, CRC prevention measures need to be conducted earlier in HIV-infected patients. In fact, some guidelines already recommend CRC screening at a lower age on the basis of data showing that CRC develops at a younger age in HIV-infected patients than in the general population and is more aggressive [19, 20, 30, 34-39]. These observations are further supported by our study findings. In addition, the advantage of colonoscopy over other screening tests is clear, as cancer precursors can instantly be identified and safely removed.
Our study has limitations. The nonrandomized, single-centre design as well as the sample size limit application of our findings to the overall population of HIV-infected individuals. However, our findings clearly suggest that further research is warranted, particularly across countries and with HIV-negative controls.
The increased prevalence of dysplastic colonic changes in the HIV-positive population in comparison with the general population confirms the value of colonoscopy as a screening test for CRC in > 50-year-old HIV-infected patients. The longer life expectancy and the elevated cancer risk in the HIV-infected population, as well as the observation that successful antiretroviral therapy does not seem to significantly affect mucosal degeneration, highlights the need for increased attention to this patient group in clinical practice. The ideal age for screening colonoscopy in HIV-infected patients has yet to be determined in large prospective studies.