The world's highest prevalence of HIV infection is found in sub-Saharan Africa, with an estimated 24.7 million adults and children older than the age of 15 years living with HIV/AIDS.1 In Zimbabwe, even though the rate of new HIV infection has been found to have declined significantly between 2003 and 2005, the United Nations Program on HIV/AIDS (UNAIDS) estimated that approximately 24% of Zimbabwean women aged 15 to 44 years visiting antenatal clinics were living with HIV in 2004.2 According to the Zimbabwean Ministry of Health and Child Welfare, however, <25,000 patients living with HIV/AIDS in Zimbabwe receive antiretroviral therapy (ART), whereas the World Health Organization (WHO) estimates that 321,000 are in need of this treatment.3 Furthermore, even if ART becomes more widely available in resource-poor countries, the high cost of measuring immunologic markers of HIV disease progression is a barrier to determining who is eligible for these treatments. The situation is further complicated by the re-emergence of tuberculosis (TB) in sub-Saharan Africa, where millions are coinfected with HIV and TB.4,5 TB prophylaxis is not routinely administered, however, because most of the population in this part of the world is purified protein derivative (PPD)-positive. Again, the prohibitive cost of immunologic markers to measure HIV disease stage is a barrier to assessing which patients should receive TB prophylaxis to prevent TB disease among individuals who are HIV-infected. Therefore, there is a serious need for markers of HIV disease progression that could play a role in assessing eligibility to ART or prophylactic regimens of specific HIV comorbidities, such as TB or even Pneumocystis jirovecii pneumonia.
Multiple studies on the prevalence of HIV-related oral lesions have been conducted in industrialized countries, before and after the availability of highly active antiretroviral therapy (HAART), and have shown that oral candidiasis (OC) is the most common HIV-related oral condition.6-9 The strong association between a high prevalence of OC and low CD4 cell count has also been widely demonstrated.10-19 Furthermore, OC has been found to decrease dramatically among populations on HAART.20-27 Studies in sub-Saharan Africa have also found OC to be the most common HIV-related oral disease.28-32 Therefore, OC may be a surrogate marker of HIV disease progression that could be used, alone or in combination with other clinical markers, to determine a starting point for ART or prophylactic regimens of specific HIV comorbidities. Although several studies among sub-Saharan African populations have shown that OC was the most common oral condition in these populations, few studies have explored OC in relation to stage of HIV disease in resource-poor settings. Furthermore, most studies on HIV-related oral diseases conducted so far in sub-Saharan Africa have relied on dental professionals to measure oral outcomes. If OC is to be used as a surrogate marker for HIV disease stage in the administration of ART or prophylactic regimens of specific HIV comorbidities, it should be measured accurately by nondental professionals, such as nurses, who are the health care professionals most frequently involved with the care and screening of HIV-infected patients in resource-limited settings. Therefore, the objectives of the present study were (1) to estimate the prevalence of oral disease, particularly OC, among Zimbabwean women in relation to HIV serostatus and CD4 cell count and (2) to assess sensitivity, specificity, and agreement between oral soft tissue examinations performed by nurses as compared with an oral surgeon.
Study Design and Participant Recruitment
We conducted a prospective study among women of reproductive age in Harare, Zimbabwe. Between November 19, 2001 and November 10, 2005, we recruited women among participants in 2 ongoing HIV-related studies of the effect of hormonal contraception on HIV transmission and HIV disease progression.33 The women were recruited from family planning clinics and outreach programs for the parent studies. They received an HIV antibody test as part of the screening procedures for the parent studies. HIV-infected and noninfected women were recruited into the present study. Participants received a free dental cleaning, basic dental care if needed, and reimbursement for transportation. Oral mucosal lesions diagnosed were also treated as per standard of care. OC was treated with nystatin oral solution or miconazole gel depending on availability. Approval to conduct this study was obtained from local ethics committees, the Medical Research Council of Zimbabwe (MRCZ), the Biomedical Research and Training Institute (BRTI), and the University of California San Francisco Committee for Human Research. Signed informed consent was obtained by nurse-examiners when women presented for one of the parent studies. The present report describes results from the baseline visit.
Variables and Measures
Outcome Variables and Oral Examination
Oral lesions investigated include OC (pseudomembranous and erythematous candidiasis and angular cheilitis), hairy leukoplakia (HL), ulcers (aphthous and recurrent herpetic ulcers), warts, Kaposi sarcoma (KS), non-Hodgkin lymphoma (NHL), and parotid enlargement. Established and standardized definitions and diagnostic criteria34,35 were used for the presumptive diagnosis of these lesions, and mucosal biopsies were conducted to confirm suspected diagnoses of malignancies (KS and NHL).
Three nurse-examiners received standardized training in the clinical diagnosis of oral mucosal lesions associated with HIV from one of the study investigators (CS). They were periodically retrained by the same investigator once a year and had access to the study manual of procedures, which included color photographs and clinical descriptors of all oral lesions investigated. The nurse-examiners performed the oral examination with the participant seated in a dental chair with a dental overhead light and recorded clinical descriptors and diagnoses on standardized forms. Each participant was then given an appointment within a 2-week window with the local Principal Investigator of the study (MMC), an oral and maxillofacial surgeon from the Department of Dentistry at the College of Health Sciences at the University of Zimbabwe. The oral surgeon attended all training sessions administered to the nurses. Clinical descriptors and diagnoses identified by the oral surgeon were also recorded on standardized forms.
Independent Variables: Laboratory Markers and Sociodemographics
HIV antibody and CD4 lymphocyte testing was performed as part of the parent studies. The CD4 cell count was measured usually on the same day as the oral study visit or, at most, within 1 month. The nurses and oral surgeon were blinded to CD4 cell count, because these results were not obtained until after the oral examination. The oral surgeon was also blinded to HIV serostatus. It was not possible to blind the nurse-examiners to HIV serostatus effectively, however, because they were also involved in the parent studies from which participants were recruited.
A standardized oral health questionnaire translated into Shona, the local language in Zimbabwe, was designed to collect sociodemographic information and oral health history and was administered to the participants at baseline.
We used proportions to summarize categorical sample characteristics and median and range to summarize continuous variables like age. We computed the prevalence of each oral lesion at baseline and compared it in HIV-infected and noninfected women using the Fisher exact test of association. We explored the prevalence of various types of OC in relation to CD4 count, categorized as <200 cells/mm3, 200 to 499 cells/mm3, and >499 cells/mm3, using the Cochran-Armitage test for trend. We computed the sensitivity and specificity of oral diagnoses made by nurse-examiners as compared with diagnoses made by the oral surgeon for pseudomembranous candidiasis, erythematous candidiasis, and any type of candidiasis in HIV-infected and uninfected women. Sensitivity was computed by dividing the number of concordant positive examinations by the number of diagnoses made by the oral surgeon (as the “gold standard”) for each type of candidiasis. Specificity was computed by dividing the number of concordant negative examinations by the number of oral examinations with a negative finding made by the oral surgeon (as the gold standard) for each type of candidiasis. We also estimated the overall concordance and the κ-statistic between nurse and oral surgeon examinations for diagnoses of OC. The concordance was computed by dividing the sum of concordant positive examinations and concordant negative examinations by 100. The formula used to compute the κ-statistic was as follows:
Among 590 women initially recruited into the study, a total of 461 received oral examinations by a nurse-examiner and the oral surgeon within a 2-week time interval and were included in the analysis performed for the present report. Nearly all the women who were approached in the parent studies agreed to participate, probably because of the incentive of receiving a dental cleaning and basic dental care. A total of 320 women were found to be HIV-infected, and 141 were found to be uninfected (Table 1). The median age was slightly higher among the HIV-infected women (30 years, ranging from 18 to 48 years) than among uninfected women (27 years, ranging from 19 to 39 years). One third (34%) of the HIV-infected study participants had a CD4 count <200 cells/mm3 when the baseline oral examination was performed, and 34% had a CD4 count ranging from 200 to 499 cells/mm3. Less than one quarter (21%) had a CD4 count >499 cells/mm3 at baseline (see Table 1). Two thirds of the study participants (64% of HIV-infected and 60% of noninfected women) had the oral examination performed by a nurse-examiner and the examination performed by the oral surgeon within 3 days of each other. Only 9% of HIV-infected and 11% of noninfected participants had the 2 examinations performed within 10 to 14 days of each other. The distribution of time intervals between examinations by the nurses and the oral surgeon did not differ by HIV status (P = 0.7).
The most common oral lesion diagnosed among HIV-infected participants by nurse-examiners and the oral surgeon was OC (22% and 25%, respectively), predominantly pseudomembranous candidiasis (16% and 14%, respectively) (Table 2). The oral surgeon found a prevalence of erythematous candidiasis almost twice as high as that of the nurse-examiners (13% vs. 7%), however. The oral surgeon also found a much higher prevalence of OC of any type among noninfected women than the nurse-examiners. Nevertheless, the nurse-examiners and oral surgeon found a significantly higher prevalence of OC among HIV-infected women than among uninfected women (P < 0.001 for nurse-examiners and P = 0.03 for the oral surgeon when comparing HIV-infected and uninfected women).
The prevalence of OC diagnosed by the nurses or the oral surgeon was significantly higher among women with a CD4 count <200 cells/mm3 than in women with a CD4 count from 200 to 499 cells/mm3 or >499 cells/mm3 (38%, 17%, and 7%, respectively; P trend < 0.001 for nurse examinations, and 49%, 17%, and 4%, respectively; P trend < 0.001 for oral surgeon examinations; Table 3). This was also found for specific types of OC (pseudomembranous or erythematous) and for nurse-examiner diagnoses and oral surgeon diagnoses.
The sensitivity of the nurses' diagnosis of any type of OC among HIV-infected women was 73%, and the specificity was 95% (Table 4). Sensitivity and specificity estimates were similar to these figures for pseudomembranous candidiasis (69% and 92%, respectively), but the sensitivity of nurses' diagnosis of erythematous candidiasis was lower (51%), whereas the specificity remained high (99%). We also compared the diagnoses of OC made by the nurses and the surgeon without designating the oral surgeon as the gold standard, using the κ-statistic. We found a κ-statistic of 0.71 for any type of OC (see Table 4). We also computed the sensitivity, specificity, and κ-statistic among HIV-infected women with a CD4 cell count <200 cells/mm3 and found similar results as those for the entire HIV-infected group.
This study is the first to explore HIV-related oral lesions in relation to CD4 cell count in a large sample of African women and the sensitivity, specificity, and agreement between diagnoses of oral lesions made by nondental professionals (nurses) and those made by an oral surgeon. Not surprisingly, we found that the most common oral lesions present at the baseline visit among HIV-infected women were pseudomembranous and erythematous candidiasis. We also found a strong statistically significant association between the presence of OC of either type and a low CD4 cell count among HIV-infected women, whether the oral examination was performed by the nurses or by the oral surgeon. The specificity of nurses' diagnoses of specific types of OC as compared with the oral surgeon's diagnoses among HIV-infected women was high (>90%), whereas the sensitivity was 73% for any type of OC. Our study results are consistent with those of previous studies conducted in developed countries and resource-poor countries, which found that the most common oral lesion associated with HIV disease is OC.10,11,13-19,28-31 Also, similar to studies conducted in the United States,10,11,13-19,36 our study showed a strong association between the occurrence of OC and a low CD4 cell count among women in sub-Saharan Africa.
The strengths of this study include a large sample of nearly 500 women recruited from family planning clinics; thus, the participating women had no oral chief complaint (they were enrolling in a study of the effect of hormonal contraception on HIV transmission and progression). Therefore, we believe that our prevalence figures are representative of the population of women who visit family planning clinics in Harare, which is likely close to the general population of women of reproductive age. Furthermore, our study population was evenly distributed by levels of CD4 cell count, which offered the opportunity to explore oral diseases in relation to a wide spectrum of HIV disease stages. We encountered an unexpected issue in that the blinding of the nurse-examiners to HIV serostatus was unsuccessful, whereas the oral surgeon was successfully blinded to that variable. This likely resulted in an underestimation of OC among HIV-uninfected women when evaluated by the nurses. Thus, we could not validly assess the interexamination agreement among HIV-uninfected women. There did not seem to be any overestimation of oral disease outcomes by nurses among HIV-infected women, however, because nurses observed OC in 22% of participants compared with the oral surgeon, who diagnosed it in 25% of participants. It is also notable that nurses diagnosed erythematous candidiasis in half the number of cases as diagnosed by the oral surgeon (23 vs. 41 cases). This is not surprising, because erythematous candidiasis exhibits discrete and subtle clinical signs at times, such as a mild patchy erythema on the dorsum of tongue and/or palate, and may easily be missed. It often occurs concomitantly with angular cheilitis or pseudomembranous candidiasis, however; thus, OC would be captured by either of these other 2 diagnoses in a case of concomitant occurrence.
The strong positive association between the occurrence of OC and a low CD4 cell count is an important result in light of the need for inexpensive surrogate markers of HIV disease progression in resource-poor countries, where measurement of CD4 cell count is prohibitively expensive. Furthermore, if OC is to be used in that context, it should be diagnosed accurately by health care professionals who are in the first line of management of patients with HIV disease, such as nurses. The good agreement between nurses and the oral surgeon with respect to the diagnosis of OC (as demonstrated by a κ-statistic of 0.71) and the moderately high sensitivity (73%) of nurses' diagnoses when the oral surgeon is used as the gold standard are good indications that nurses can reliably diagnose OC in HIV-infected patients. A visual inspection of the mouth is quick (<5 minutes), noninvasive, and inexpensive, and it may likely be used in combination with other clinical indicators in the decision-making process of when to initiate ART or prophylactic regimens of specific HIV comorbidities.
The authors express their gratitude to the study nurse-examiners, Letticia Chidzumo and Edelyn Makura, who performed the oral examinations and to all study participants. They also thank Dr. Deborah Greenspan for her valuable input with respect to the initial study design; Drs. Tsungai Chipato, Mike Chirenje, Charles Morrison, and Nancy Padian for allowing them access to participants in the hormonal contraception studies; and Dr. Art Reingold for facilitating the initial funding for this project.
4. Dye C, Scheele S, Dolin P, et al. Consensus statement. Global burden of tuberculosis: estimated incidence, prevalence, and mortality by country. WHO Global Surveillance and Monitoring Project. JAMA
5. Bleed D, Dye C, Raviglione MC. Dynamics and control of the global tuberculosis epidemic. Curr Opin Pulm Med
6. Shiboski CH. Epidemiology of HIV
-related oral manifestations in women
: a review. Oral Dis
7. Shiboski CH. HIV
-related oral disease epidemiology among women
: year 2000 update. Oral Dis
8. Coogan MM, Greenspan J, Challacombe SJ. Oral lesions in infection with human immunodeficiency virus. Bull World Health Organ
9. Hodgson TA, Greenspan D, Greenspan JS. Oral lesions of HIV
disease and HAART in industrialized countries. Adv Dent Res
10. Feigal DW, Katz MH, Greenspan D, et al. The prevalence of oral lesions in HIV
-infected homosexual and bisexual men: three San Francisco epidemiological cohorts. AIDS
11. Barr CE, Lopez MR, Rua-Dobles A, et al. HIV
-associated oral lesions; immunologic, virologic and salivary parameters. J Oral Pathol Med
12. Katz MH, Greenspan D, Westenhouse J, et al. Progression to AIDS in HIV
-infected homosexual and bisexual men with hairy leukoplakia and oral candidiasis
13. Shiboski CH, Hilton JF, Greenspan D, et al. HIV
-related oral manifestations in two cohorts of women
in San Francisco. J Acquir Immune Defic Syndr Hum Retrovirol
14. Glick M, Muzyka BC, Lurie D, et al. Oral manifestations associated with HIV
-related disease as markers for immune suppression and AIDS. Oral Surg Oral Med Oral Pathol
15. Shiboski CH, Hilton JF, Neuhaus JM, et al. Human immunodeficiency virus-related oral manifestations and gender. A longitudinal analysis. Arch Intern Med
16. Schuman P, Ohmit SE, Sobel JD, et al. Oral lesions among women
living with or at risk for HIV
Epidemiology Research Study (HERS) Group. Am J Med
17. Shiboski CH, Neuhaus JM, Greenspan D, et al. The effect of receptive oral sex and smoking on the incidence of hairy leukoplakia in HIV
-positive gay men. J Acquir Immune Defic Syndr
18. Patton L. Sensitivity, specificity, and positive predictive value of oral opportunistic infections in adults with HIV
/AIDS as markers of immune suppression and viral burden. Oral Surg Oral Med Oral Pathol Oral Radiol Endod
19. Greenspan D, Komaroff E, Redford M, et al. Oral mucosal lesions and HIV
viral load in the Women
's Interagency HIV
Study. J Acquir Immune Defic Syndr
20. Arribas J, Hernandez-Albujar S, Gonzalez-Garcia J, et al. Impact of protease inhibitor therapy on HIV
-related oropharyngeal candidiasis. AIDS
21. Patton L, McKaig R, Strauss R, et al. Changing prevalence of oral manifestations of human immuno-deficiency virus in the era of protease inhibitor therapy. Oral Surg Oral Med Oral Pathol Oral Radiol Endod
22. Schmidt-Westhausen A, Priepke F, Bergmann F, et al. Decline in the rate of oral opportunistic infections following introduction of highly active antiretroviral therapy. J Oral Pathol Med
23. Tappuni AR, Fleming GJ. The effect of antiretroviral therapy on the prevalence of oral manifestations in HIV
-infected patients: a UK study. Oral Surg Oral Med Oral Pathol Oral Radiol Endod
24. Ramirez-Amador V, Esquivel-Pedraza L, Sierra-Madero J, et al. The changing clinical spectrum of human immunodeficiency virus (HIV
)-related oral lesions in 1000 consecutive patients: a 12-year study in a referral center in Mexico. Medicine
25. Greenspan D, Gange S, Phelan J, et al. Incidence of oral lesions in HIV
: reduction with HAART. J Dent Res
26. Nicolatou-Galitis O, Velegraki A, Paikos S, et al. Effect of PI-HAART on the prevalence of oral lesions in HIV
-I infected patients. A Greek study. Oral Dis
27. Umadevi KM, Ranganathan K, Pavithra S, et al. Oral lesions among persons with HIV
disease with and without highly active antiretroviral therapy in southern India. J Oral Pathol Med
28. Hodgson TA. HIV
-associated oral lesions: prevalence in Zambia. Oral Dis
. 1997;3(Suppl 1):S46-S50.
29. Arendorf T, Bredekamp B, Cloete C, et al. Oral manifestations of HIV
infection in 600 South African patients. J Oral Pathol Med
30. Matee M, Scheutz F, Moshy J. Occurrence of oral lesions in relation to clinical and immunological status among HIV
-infected adult Tanzanians. Oral Dis
31. Chidzonga M. HIV
/AIDS orofacial lesions in 156 Zimbabwean patients at referral oral and maxillofacial surgical clinics. Oral Dis
32. Ranganathan K, Hemalatha R. Oral lesions in HIV
infection in developing countries: an overview. Adv Dent Res
33. Morrison CS, Richardson BA, Mmiro F, et al. Hormonal contraception and the risk of HIV
34. Greenspan JS, Barr CE, Sciubba JJ, et al. Oral manifestations of HIV
infection. Definitions, diagnostic criteria, and principles of therapy. The U.S.A. Oral AIDS Collaborative Group. Oral Surg Oral Med Oral Pathol Oral Radiol Endod
35. EC-Clearinghouse. Classification and diagnostic criteria for oral lesions in HIV
infection. EC-Clearinghouse on Oral Problems Related to HIV
Infection and WHO Collaborating Centre on Oral Manifestations of the Immunodeficiency Virus. J Oral Pathol Med
36. Selwyn PA, Alcabes P, Hartel D, et al. Clinical manifestations and predictors of disease progression in drug users with human immunodeficiency virus infection. N Engl J Med