| |
Psychosocial, Medical, and Gynecologic Complications in a Cohort of HIV-Infected Women From an Underserved Minority Community in Los Angeles, California
|
| |
| |
JAIDS Journal of Acquired Immune Deficiency Syndromes: Volume 41(2) 1 February 2006 pp 255-257
LETTER TO THE EDITOR
Ogunyemi, Oreoluwa MS; Ajayi, Olusegun MD; Rodriguez, Noah MD
Department of Obstetrics and Gynecology, David Geffen School of Medicine, University of California, Los Angeles (UCLA), CA, Harbor-UCLA Medical Center, Torrance, CA
Presented at the European Board and College of Obstetrics and Gynecology XVIIIth European Congress on Obstetrics and Gynecology; 2004; Athens.
We studied the ethnic, psychosocial, and gynecologic associations of HIV infections in a minority population. This was a retrospective analysis of 125 HIV-infected women, including 36 Hispanics (29%), 86 African Americans (69%), and 3 others (2%), managed in an underserved community in Los Angeles between 1982 and 2002. Institutional Review Board approval was obtained. The subjects' medical data were then divided into pre-HIV and post-HIV medical complications. Statistical analysis was done as indicated using the commercially available SPSS statistical program. P values less than 0.05 and a 95% confidence interval (CI) that did not include 1.00 were considered statistically significant.
Alcohol abuse occurred in 33 (26%) women, tobacco use in 36 (29%), cocaine abuse in 29 (23%), incarceration in a correctional facility in 19 (15%), partner violence in 13 (10%), and prostitution in 6 (5%). Thirty (24%) women had a known HIV-positive partner, and 16 (13%) were pregnant at the time of diagnosis. The source of HIV infection was unprotected intercourse in 75 (60%) women, intravenous drugs in 28 (22%), blood transfusion in 8 (6%), and unknown in 14 (10%). As expected, HIV opportunistic infections occurred in 1 (0.8%) woman before and in 35 (28%) women after HIV diagnosis (odds ratio [OR] = 48). Psychiatric disease occurred in 14% versus 32% after HIV diagnosis. Surgery showed a likelihood ratio of a 50% decrease with onset of HIV infections.
Cervical dysplasia occurred in 14% after HIV diagnosis versus 2% before diagnosis (P = 0.001; OR = 6.9). There was an increased risk of infections with candidiasis (0.8% vs. 16.8%; P = 0.000; OR = 25); trichomoniasis (0.8% vs. 10%; P = 0.005; OR = 10.8), and human papillomavirus (2% vs. 8%; P = 0.46) after the diagnosis of HIV but not with bacterial vaginosis. None of the sexually transmitted diseases (STDs; herpes, Chlamydia, syphilis, and gonorrhea) showed any significant differences before or after HIV infections. The need for gynecologic surgeries, including hysterectomies, was not affected by HIV diagnosis in this group of women. However, menstrual disorders occurred only in these women after they were diagnosed with HIV (0% vs. 10.4%; P = 0.000; 95% CI: 1.1 to 1.2). Because it was expected that menstrual disorders would increase with age, correlation analysis was performed with age of diagnosis with HIV, duration of follow-up in the clinic, and last age at follow-up, and there was no significant correlation between menstrual disorders and any of these age-related variables.
A comparison was also performed between African American and Hispanic HIV-infected women to assess possible ethnic differences in HIV presentation (Table 1). There was no difference between the 2 groups in the mean age of diagnosis; however, African American women compared with Hispanic women were significantly more likely to abuse alcohol, cocaine, and tobacco, with no difference in the 2 groups with intravenous drug abuse (IVDA). African American women were also more significantly incarcerated and exposed to partner violence compared with Hispanic women with HIV. Even though there were no statistical differences between the 2 groups for prostitution, all the women who admitted to prostitution were African American. Interestingly, Hispanic women significantly reported unprotected intercourse as a source of infection and were more likely to be diagnosed with HIV during pregnancy. There were no statistical differences between the 2 groups regarding homelessness or HIV-infected partners. There were also no significant differences between the 2 groups regarding nongynecologic infections or psychiatric disorders; however, African Americans had significantly more medical comorbidity. There were no statistical differences between the 2 groups for viral loads and CD4 cell counts, but African American women had a statistically prolonged interval of approximately 18 more months from diagnosis to achieving the lowest viral load. For gynecologic disorders, trichomoniasis and candidiasis occurred significantly more often in African American women and menstrual disorders occurred significantly more often in Hispanic women. There were no significant differences between the 2 groups in relation to any of the STDs, human papillomavirus (HPV), or cervical dysplasia.
TABLE 1. Comparison of Complications in a Cohort of African-American Versus Hispanic Women Infected With HIV
Correlates of cervical dysplasia were performed with all variables of gynecologic, psychosocial, and medical complications. This showed that the strongest correlation with cervical dysplasia was with HPV infections (r = 0.514, P = 0.000). There was a correlation with all the categories of medical infections, with the strongest correlation occurring with HIV opportunistic infections (r = 0.245, P = 0.000). There was also a positive correlation with all STDs and vaginitis, but there were only individual correlations with candidiasis, Chlamydia, and syphilis. Cervical dysplasia showed a positive correlation with highest viral loads (r = 0.218, P = 0.017) and a negative correlation with lowest CD4 counts (r = -0.247, P = 0.006). There was also a positive correlation with the interval from diagnosis to initiation of highly active antiretroviral therapy (HAART) (r = 0.264, P = 0.009). There was no correlation with HAART use.
A linear regression analysis was done to determine the independent variables predictive of cervical dysplasia. The only variables that were independently predictive of cervical dysplasia were viral load at diagnosis (ƒÀ = 0.412, P = 0.001), HPV (ƒÀ = 0.163, P = 0.04), and interval from diagnosis to highest viral load (ƒÀ = 0.202, P = 0.007). The model summary showed an R of 0.432, and the F change significance was 0.000, suggesting an approximately 50% significant linear relation between the model and cervical dysplasia. The sum of squares for the regression was 1.083, and the residual sum of squares was 4.711, suggesting that this model explained approximately 22% of the variation for cervical dysplasia presentation. HAART use was not included in the regression analysis because all the women with cervical dysplasia eventually received HAART.
Hispanic women had a significant association with unprotected sexual intercourse and HIV diagnosis during pregnancy, whereas infected African American women had an association with substance abuse, violence, incarceration, and probably commercial sex. This suggests that Hispanic women may be exposed to HIV in the traditional family setting with a partner and children, whereas African American women are more likely to be exposed to HIV in a dysfunctional social setting. This information may be useful in developing intervention prevention strategies for the target population.
Consistent with previous studies, HIV diagnosis was associated with cervical dysplasia. Previous studies all agree with the role of CD4 counts, but the relation to viral load has been inconsistent.1,2 This study supports the importance of viral load by showing that highest mean viral load correlated positively with cervical dysplasia, whereas viral load at diagnosis and interval from diagnosis to highest viral load were significant independent predictors of cervical dysplasia. Previous studies have suggested that the effect of HAART on cervical dysplasia has not been established. The results from this study suggest that the interval to initiation of HARRT may be important in predicting cervical dysplasia, because most patients eventually receive HAART. An interesting finding is the increase in menstrual disorders with HIV diagnosis, which was not associated with age or duration of disease, and this agrees with the findings of Minkoff et al.3 The reasons for an association with menstrual disorders are not clear but may be related to stress associated with HIV infections, side effects of medication, or even a direct effect of HIV infection.
African American women were more likely to have trichomoniasis and candidiasis than Hispanic HIV-infected women, with no difference in STD or cervical dysplasia frequencies. Hispanic women had more menstrual disorders, however. African American women had more medical comorbidity, which may relate to the health disparity burden associated with African American ethnicity. The significantly prolonged time to lowest viral load in African American women may signify compliance differences with ethnicity. Ethnic differences in the health burden of women infected with HIV may be useful in health care planning and policy, and this area deserves further assessment.
In conclusion, there are ethnic variations in psychosocial and gynecologic presentations of HIV-infected women. Immunosuppression as measured by CD4 cell counts and viral load was predictive of cervical dysplasia in this cohort.
Dotun Ogunyemi, MD*
Oreoluwa Ogunyemi, MS*
Olusegun Ajayi, MD
Noah Rodriguez, MD*
*David Geffen School of Medicine University of California Los Angeles (UCLA), CA, Harbor-UCLA Medical Center Torrance, CA
|
|
| |
| |
|
|
|
