GRADE tables

Freeman F, Maurer T, Gbabe O, Okwundu CI, Laker M, Easterbrook PE, et al. Treatment for mild and moderate Kaposi’s sarcoma in ART-naive HIV-infected individuals. Cochrane Database Syst Rev. (in press)
Mosam A, Shaik F, Uldrick TS, Esterhuizen T, Friedland GH, Scadden DT, et al. A randomized controlled trial of HAART versus HAART and chemotherapy in therapy-naïve patients with HIV-associated Kaposi sarcoma in South Africa. J Aquir Immune Defic Syndr. 2012;60(2):150–7. [PMC free article: PMC3360837] [PubMed: 22395672]

Tinea infections

Bagatell FK. A prospective study of bifonazole 1% cream in the once-daily management of tinea corporis/cruris. Adv Therapy. 1986;3(5):294–300.
Dobson RL, Bagatell FK, Hickman JG, Whitmore CG, Willis I, Seffon J, et al. Naftifine 1% cream in the treatment of tinea cruris and tinea corporis. Drug Invest. 1991;3(1):57–9.
Hantschke D, Reichenberger M. Doppelblinde, randomisierte vergleichende in vivo Untersuchungen zwischen den Antimykotika Clotrimazol, Tolnaftat und Naftifin. [Double blind, randomized in vivo investigations comparing the antifungals clotrimazole, tolnaftate and naftifine (author’s transl)]. Mykosen. 1980;23(12):657–68. (in German) [PubMed: 7012610]
Haroon TS, Hussain I, Aman S, Jahangir M. Randomized, comparative, study of 1% naftifine cream (once daily) and 1% tioconazole cream (twice daily) in the treatment of tinea cruris. Pakistan Journal of Medical Sciences Quarterly. 1996;12(2):181–4.
Jerajani HR, Janaki C, Kumar S, Phiske M. Comparative assessment of the efficacy and safety of sertaconazole (2%) cream versus terbinafine (1%) cream versus luliconazole (1%) cream in patients with dermatophytoses: a pilot study. Indian J Dermatol. 2013;58(1):34–8. [PMC free article: PMC3555370] [PubMed: 23372210]
Katz HI, Bard J, Cole GW, Fischer S, McCormick GE, Medansky RS, et al. SCH 370 (clotrimazole-betamethasone dipropionate) cream in patients with tinea cruris or tinea corporis. Cutis. 1984;34(2):183–8. [PubMed: 6383734]
Lebwohl M, Elewski B, Eisen D, Savin RC. Efficacy and safety of terbinafine 1% solution in the treatment of interdigital tinea pedis and tinea corporis or tinea cruris. Cutis. 2001;67(3):261–6. [PubMed: 11270304]
Li M, Bi ZG, Gu J, Sheng YN, Zhang MH, Wang Y, et al. Clinical study of butenafine hydrochloride 1% cream in the treatment of tinea pedis, tinea corporis and tinea cruris. J Clinl Dermatol. 2006;35(7):471–2.
Millikan LE. Efficacy and tolerability of topical terbinafine in the treatment of tinea cruris. J Am Acad Dermatol. 1990;23(4.2):795–9. [PubMed: 2229526]
Miura Y, Onuki M, Takahashi S, Seiji M, Sato A, Kagawa S, et al. A double-blind study on utility of econazole cream in dermatomycosis. Rinsho Hyoka (Clinical Evaluation). 1979;7(1):83–108.
Pariser RJ, Pariser DM. Clinical and mycological effect of clotrimazole/betamethasone dipropionate cream versus ketoconazole cream in patients with tinea cruris. J Dermatol Treat. 1995;6(3):173–7.
Ramam M, Prasas HR, Manchanda Y, Khaitan BK, Banerjee U, Mukhopadhyaya A, et al. Randomised controlled trial of topical butenafine in tinea cruris and tinea corporis. Indian J Dermatol Venereol Leprol. 2003;69(2):154–8. [PubMed: 17642865]
Shen WM, Hu YW, Gu HY. Econazole compound cream vs miconazole cream in treating tinea corporis & cruris. Chinese Journal of New Drugs and Clinical Remedies. 2002;21(3):143–5.
Singal A, Pandhi D, Agrawal S, Das S. Comparative efficacy of topical 1% butenafine and 1% clotrimazole in tinea cruris and tinea corporis: a randomized, double-blind trial. J Dermatol Treatment. 2005;16(5–6):331–5. [PubMed: 16428155]
Spiekermann PH, Young MD. Clinical evaluation of clotrimazole, a broad-spectrum antifungal agent. Arch Dermatol. 1976;112(3):350–2. [PubMed: 769697]
Tanenbaum L, Taplin D, Lavelle C, Akers WA, Rosenberg MJ, Carmargo G. Sulconazole nitrate cream 1 percent for treating tinea cruris and corporis. Cutis. 1989;44(4):344–7. [PubMed: 2805811]
Wang AP, Li RY, Shun QN, Wan Z, Wang XH, Wang JB, et al. A double blind randomized controlled clinical trial of econazole-triamcinolon acetonide cream in the treatment of tinea pedis and tinea corporis & cruris. The Chinese Journal of Clinical Pharmacology. 2000;16(5):345–9.
Wortzel MH. A double-blind study comparing the superiority of a combination antifungal (clotrimazole)/steroidal (betamethasone dipropionate) product. Cutis. 1982;30(2):258–61. [PubMed: 6751707]
Zaias N, Berman B, Cordero CN, Hernandez A, Jacobson C, Millikan L, et al. Efficacy of a 1-week, once-daily regimen of terbinafine 1% cream in the treatment of tinea cruris and tinea corporis. J Am Acad Dermatol. 1993;29(4):646–8. [PubMed: 8408803]

Oropharyngeal candidiasis

De Wit D, O’Doherty E, De Vroey C, Clumneck N. Safety and efficacy of single-dose fluconazole compared with a 7-day regimen of itraconazole in the treatment of AIDS-related oropharyngeal candidiasis. J Int Med Res. 1998;26(3):159–70. [PubMed: 9718471]
Linpiyawan R, Jittreprasert K, Sivayathorn A. Clinical trial: clotrimazole troche vs. itraconazole oral solution in the treatment of oral candidosis in AIDS patients. Int J Dermatol. 2000;39(11):859–61. [PubMed: 11123451]

Footnotes

1: This list only includes references cited in footnotes. The studies included in the GRADE tables are listed in the systematic reviews.

Publication Details

Copyright

All rights reserved. Publications of the World Health Organization are available on the WHO website (www.who.int) or can be purchased from WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791 4857; e-mail: tni.ohw@sredrokoob). Requests for permission to reproduce or translate WHO publications – whether for sale or for non-commercial distribution – should be addressed to WHO Press through the WHO website (www.who.int/about/licensing/copyright_form/en/index.html).

Publisher

World Health Organization, Geneva

NLM Citation

Guidelines on the Treatment of Skin and Oral HIV-Associated Conditions in Children and Adults. Geneva: World Health Organization; 2014. Web Appendix 1, GRADE tables.

Table 1.1Should chemotherapy plus ART versus ART alone be used for Kaposi sarcoma for children with HIV infection?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Chemo + ART	ART	Relative (95% CI)	Absolute	Quality	Importance
Complete response to treatment
1	Observational study	Very serious¹^,²	No serious inconsistency	No serious indirectness	Very serious³	None	17/26 (65.4%)	2/13 (15.4%)	RR 4.25 (1.15 to 15.68)	500 more per 1000 (from 23 more to 1000 more)	⊕○○○ VERY LOW	CRITICAL
Complete/partial response to treatment
1	Observational study	Very serious¹^,²	No serious inconsistency	No serious indirectness	Very serious³	None	23/26 (88.5%)	2/13 (15.4%)	RR 5.75 (1.59 to 20.73)	731 more per 1000 (from 91 more to 1000 more)	⊕○○○ VERY LOW	CRITICAL
Complete among known outcome
1	Observational study	Very serious¹	No serious inconsistency	No serious indirectness	Very serious³	None	17/24 (70.8%)	2/2 (100%)	RR 0.84 (0.48 to 1.48)	160 fewer per 1000 (from 520 fewer to 480 more)	⊕○○○ VERY LOW	CRITICAL
Complete/partial among known outcome
1	Observational study	Very serious¹	No serious inconsistency	No serious indirectness	Very serious³	None	23/24 (95.8%)	2/2 (100%)	RR 1.13 (0.67 to 1.89)	130 more per 1000 (from 330 fewer to 890 more)	⊕○○○ VERY LOW	CRITICAL
Mean CD4% increase during chemotherapy (better indicated by higher values)
1	Observational study	Serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	None	24	24	—	MD 13.2 higher (1.65 to 24.65 higher)	⊕○○○ VERY LOW	CRITICAL
Mortality
1	Observational study	Very serious¹	No serious inconsistency	No serious indirectness	Very serious³	None	13/36 (36.1%)	7/14 (50%)	RR 0.72 (0.37 to 1.43)	140 fewer per 1000 (from 315 fewer to 215 more	⊕○○○ VERY LOW	CRITICAL

1: Unadjusted estimates.

2: Many patients had missing outcome data.

3: Very few cases (<50).

Table 1.2Should chemotherapy plus ART versus ART alone be used for Kaposi sarcoma for children with HIV infection?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Chemo + ART	ART	Relative (95% CI)	Absolute	Quality	Importance
Complete response to treatment
1	Observational study	Very serious¹^,²	No serious inconsistency	No serious indirectness	Very serious³	None	17/26 (65.4%)	1/10 (10%)	RR 6.54 (1 to 42.86)	554 more per 1000 (from 0 more to 1000 more)	⊕○○○ VERY LOW	CRITICAL
Complete/partial response to treatment
1	Observational study	Very serious¹^,²	No serious inconsistency	No serious indirectness	Very serious³	None	23/26 (88.5%)	6/10 (60%)	RR 1.47 (0.87 to 2.49)	282 more per 1000 (from 78 fewer to 894 more)	⊕○○○ VERY LOW	CRITICAL
Complete among known outcome
1	Observational study	Very serious¹	No serious inconsistency	No serious indirectness	Very serious³	None	17/24 (70.8%)	1/6 (16.7%)	RR 4.25 (0.7 to 25.91)	542 more per 1000 (from 50 fewer to 1000 more)	⊕○○○ VERY LOW	CRITICAL
Complete/partial among known outcome
1	Observational study	Very serious¹	No serious inconsistency	No serious indirectness	Very serious3	None	23/24 (95.8%)	6/6 (100%)	RR 1.01 (0.81 to 1.27)	10 more per 1000 (from 190 fewer to 270 more)	⊕○○○ VERY LOW	CRITICAL
Mortality
2	Observational studies	Very serious¹	No serious inconsistency	No serious indirectness	Very serious³	None	25/66 (37.9%)⁴	22/32 (68.8%)⁴	RR 0.46 (0.29 to 0.73)	371 fewer per 1000 (from 186 fewer to 488 fewer)	⊕○○○ VERY LOW	CRITICAL

1: Unadjusted estimates.

2: Many patients had missing outcome data.

3: Very few cases (<50).

4: Imputed data from information in text for one study. Data not used in calculating RR.

Table 1.3Should highly-active antiretroviral therapy (HAART) plus ABV versus HAART alone be used for mild and moderate treatment-naive Kaposi sarcoma?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	HAART plus ABV	HAART alone	Relative (95% CI)	Absolute	Quality	Importance
Mortality (follow-up 12 months)
1	Randomized trial	No serious risk of bias	No serious inconsistency	Serious¹	Serious²	None	0/3 (0%)	1/9 (11.1%)	RR 0.83 (0.04 to 16.46)	19 fewer per 1000 (from 107 fewer to 1000 more)	⊕⊕○○ LOW	CRITICAL
Complete response (follow-up 12 months)
1³	Randomized trial	No serious risk of bias	No serious inconsistency	Serious¹	Serious²	None	1/3 (33.3%)	0/9 (0%)	RR 7.5 (0.38 to 148.13)	—	⊕⊕○○ LOW	CRITICAL
Partial response (follow-up 12 months)
1³	Randomized trial	No serious risk of bias	No serious inconsistency	Serious¹	Serious²	None	2/3 (66.7%)	5/9 (55.6%)	RR 1.2 (0.45 to 3.23)	111 more per 1000 (from 306 fewer to 1000 more)	⊕⊕○○ LOW	CRITICAL
Progression (at 12 months)
1³	Randomized trial	No serious risk of bias	No serious inconsistency	Serious¹	Serious²	None	0/3 (0%)	1/9 (11.1%)	RR 0.83 (0.04 to 16.46)	19 fewer per 1000 (from 107 fewer to 1000 more)	⊕⊕○○ LOW	CRITICAL
Stable disease (follow-up 12 months)
1³	Randomized trial	No serious risk of bias	No serious inconsistency	Serious¹	Serious²	None	0/3 (0%)	2/9 (22.2%)	RR 0.5 (0.03 to 8.27)	111 fewer per 1000 (from 216 fewer to 1000 more)	⊕⊕○○ LOW	IMPORTANT
KS IRIS
1	Randomized trial	No serious risk of bias	No serious inconsistency¹	Serious¹	Serious²	None	0/3 (0%)	0/9 (0%)	Not pooled	Not pooled	⊕⊕○○ LOW	IMPORTANT
Adverse events (Grade III-V)
1	Randomized trial	No serious risk of bias	No serious inconsistency	Serious¹	Serious²	None	0/3 (0%)	0/9 (0%)	Not pooled	Not pooled	⊕⊕○○ LOW	IMPORTANT

1: Post-hoc analysis of study not specifically designed to evaluate patients with mild to moderate disease.

2: Single study, very small post-hoc analysis of only 12 patients with mild-moderate KS.

3: See TABLE 1 in systematic review (Freeman et al., in press) for detailled definitions of outcomes. Mosam and colleagues (2012) modified these slightly, as below. Complete response (CR): resolution of any detectable disease for at least 4 weeks. Partial response (PR) is a 50% or > decrease in number and/or size of all existing lesions for at least 4 weeks, without the appearance of new lesions. A response may be assigned to a diminution in the diameter of all lesions, or to flattening of at least 50% of the lesions. The size of each lesion will be the product of the longest dimension and the maximum dimension perpendicular to it. Overall response: PR + CR. Stable disease: not meeting the criteria for progression, PR or CR. Progressive disease: at least a 25% increase in the size of any lesion or the appearance of any new lesions.

Table 1.4Should HAART plus ABV versus HAART alone be used for mild and moderate treatment-naive Kaposi sarcoma (observational studies)?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	HAART plus ABV	HAART alone	Relative (95% CI)	Absolute	Quality	Importance
Overall response (complete + partial) (follow-up mean 12 months)
1	Observational study	No serious risk of bias	No serious inconsistency	No serious indirectness	Very serious¹	None	1/1 (100%)	0/2 (0%)	RR 4.5 (0.32 to 63.94)	—	⊕○○○ VERY LOW	CRITICAL

1: Single study, very small, post-hoc analysis of only 3 participants.

Table 1.5Should HAART plus liposomal anthracyclines versus HAART alone be used for mild and moderate treatment-naive Kaposi sarcoma (RCTs)?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	HAART plus liposomal anthracyclines	HAART alone	Relative (95% CI)	Absolute	Quality	Importance
Overall response (complete + partial) (follow-up 48 weeks)
1	Randomized trial	No serious risk of bias	No serious inconsistency	Serious¹	Serious²	None	6/8 (75%)	2/8 (25%)	RR 3 (0.85 to 10.63)	500 more per 1000 (from 37 fewer to 1000 more)	⊕⊕○○ LOW	CRITICAL
Adverse events (follow-up 48 weeks)
1	Randomized trial	Serious³	No serious inconsistency	Serious¹	Serious²	None	5/14 (35.7%)	0/8 (0%)	OR 9.84 (0.47 to 205.62)	—	⊕○○○ VERY LOW	IMPORTANT

1: Post-hoc analysis of subgroup of patients with mild to moderate disease.

2: Single study of only 16 patients.

3: Adverse events only reported for intervention arm (PLD+ART), not for comparison arm (ART alone).

Table 1.6Should HAART plus liposomal anthracyclines versus HAART alone be used for mild and moderate treatment-naive Kaposi sarcoma (observational studies)?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	HAART plus liposomal anthracyclines	HAART alone	Relative (95% CI)	Absolute	Quality	Importance
Mortality¹ (follow-up 12 months)
1	Observational study	No serious risk of bias	No serious inconsistency	Serious²	Serious³	None	0/7 (0%)	0/77 (0%)	Not pooled¹	Not pooled¹	⊕○○○ VERY LOW	CRITICAL
Mortality¹ (follow-up 3 months)
1	Observational study	No serious risk of bias	No serious inconsistency	Serious²	Serious³	None	0/7 (0%)	0/77 (0%)	Not pooled¹	Not pooled¹	⊕○○○ VERY LOW	CRITICAL
KS IRIS (follow-up median 40.5 months)
1	Observational study	No serious risk of bias	No serious inconsistency	Serious²	Serious³	None	0/7 (0%)	6/71 (8.5%)	RR 0.69 (0.04 to 11.18)	26 fewer per 1000 (from 81 fewer to 860 more)	⊕○○○ VERY LOW	IMPORTANT

1: No deaths in either group.

2: Single study, where intervention group (liposomal anthracyclines plus ART) was not representative of patients with mild/moderate disease in general; UK cohort participants with mild/moderate disease only received liposomal anthracyclines above and beyond ART in exceptional circumstances per clinician’s decision (standard of care at that site was considered to be ART alone).

3: Very few patients in intervention group (7) due to standard of care in UK cohort (see ^{footnote above}).

Table 1.7Should HAART plus ABV versus HAART alone be used for severe or progressive Kaposi sarcoma?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	HAART+ABV	HAART alone	Relative (95% CI)	Absolute	Quality	Importance
Mortality
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious¹	None	11/50 (22%)	13/50 (26%)	RR 0.92 (0.45 to 1.88)	21 fewer per 1000 (from 143 fewer to 229 more)	⊕⊕⊕○ MODERATE	CRITICAL
Progressive disease (follow-up mean 12 months)
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious¹	None	1/50 (2%)	10/50 (20%)	RR 0.1 (0.01 to 0.75)	180 fewer per 1000 (from 50 fewer to 198 fewer)	⊕⊕⊕○ MODERATE	CRITICAL
Clinical response – complete response (follow-up mean 12 months)
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Very serious¹	None	8/50 (16%)	4/50 (8%)	RR 2 (0.64 to 6.22)	80 more per 1000 (from 29 fewer to 418 more)	⊕⊕○○ LOW	CRITICAL
Clinical response – partial response (follow-up mean 12 months)
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious¹	None	24/50 (48%)	14/50 (28%)	RR 1.71 (1.01 to 2.91)	199 more per 1000 (from 3 more to 535 more)	⊕⊕⊕○ MODERATE	CRITICAL

1: There were very few events with very wide CIs.

Table 1.8Should HAART + ABV versus HAART alone be used for severe or progressive Kaposi sarcoma?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	HAART+ABV	HAART alone	Relative (95% CI)	Absolute	Quality	Importance
Clinical response – stable disease (follow-up mean 12 months)
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Very serious¹	none	0/50 (0%)	8/50 (16%)	RR 0.06 (0 to 0.99)	150 fewer per 1000 (from 2 fewer to 160 fewer)	⊕⊕○○ LOW	CRITICAL
Clinical response – overall response (complete and partial) (follow-up mean 12 months)
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious¹	None	32/50 (64%)	18/50 (36%)	RR 1.78 (1.16 to 2.72)	281 more per 1000 (from 58 more to 619 more)	⊕⊕⊕○ MODERATE	CRITICAL
Adverse events (follow-up mean 12 months)
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious¹	None	23/50 (46%)	26/50 (52%)	RR 0.88 (0.59 to 1.32)	62 fewer per 1000 (from 213 fewer to 166 more)	⊕⊕⊕○ MODERATE	IMPORTANT

1: There were very few events with very wide CIs.

Table 1.9Should HAART + PLD versus HAART alone be used for severe or progressive Kaposi sarcoma?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	HAART + PLD	HAART alone	Relative (95% CI)	Absolute	Quality	Importance
Clinical response (follow-up mean 48 weeks)
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Very serious¹	None	4/5 (80%)	0/5 (0%)	RR 9 (0.61 to 133.08)	—	⊕⊕○○ LOW	IMPORTANT

1: There were very few events with very wide CIs.

Table 1.10Should HAART + liposomal anthracycline versus HAART alone be used for severe or progressive Kaposi sarcoma?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	HAART + liposomal anthracycline	HAART alone	Relative (95% CI)	Absolute	Quality	Importance
Mortality (follow-up median 4 years)
1	Observational study	No serious risk of bias	No serious inconsistency	No serious indirectness¹	Serious¹	None	5/65 (7.7%)	4/64 (6.3%)	RR 1.23 (0.35 to 4.38)	14 more per 1000 (from 41 fewer to 211 more)	⊕○○○ VERY LOW	CRITICAL
KS IRIS (follow-up median 4 years)
1	Observational study	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious¹	None	4/65 (6.2%)	8/64 (12.5%)	RR 0.49 (0.16 to 1.55)	64 fewer per 1000 (from 105 fewer to 69 more)	⊕○○○ VERY LOW	IMPORTANT

1: There were very few events with very wide CIs.

Table 2.1Should lithium succinate versus placebo be used for seborrhoeic dermatitis in HIV-infected patients?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	No lithium succinate	Lithium succinate	Relative (95% CI)	Absolute	Quality	Importance
Incidence assessed by clinical examination (follow-up 47 days)
1	Randomized Trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious imprecision¹	Undetected	10	10	—	—	⊕⊕⊕○ MODERATE¹	MODERATE

1: Difference between the intervention and placebo groups could be observed only up to 6.8 days. Out of 10 subjects recruited, 9 were there until 6.8 days and only 5 until 47 days which was the maximum follow-up reported.

Table 2.2Should pimecrolimus be used for seborrhoeic dermatitis in HIV-infected patients?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	No pimecrolimus	Pimecrolimus	Relative (95% CI)	Absolute	Quality	Importance
Resolution (important outcome; assessed with clinical exam; follow-up 4 weeks)
1	Observational study	Very serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	Undetected	—	19/19 (100%)	—	—	⊕○○○ VERY LOW¹	MODERATE
Relapse (not important outcome; follow-up 4 weeks)
1	Observational study	Very serious	No serious inconsistency	No serious indirectness	No serious imprecision	Undetected	—	2/19 (10.5%)	—	—	⊕○○○ VERY LOW¹	MODERATE

1: Open label single group pilot study.

Table 2.3Should bifonazole be used for seborrhoeic dermatitis in HIV-infected patients?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	No bifonazole	Bifonazole	Relative (95% CI)	Absolute	Quality	Importance
Resolution (important outcome; assessed with clinical exam; follow-up 4 weeks)
1	Observational study	Serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	Undetected	—	12/15 (80%)	—	—	⊕⊕○○ LOW	MODERATE
Relapse (not important outcome; follow-up 4 weeks)
1	Observational study	Serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	Undetected	—	9/12 (75%)	—	—	⊕○○○ VERY LOW¹	MODERATE

1: No explanation was provided.

Table 2.4Should ART be used for seborrhoeic dermatitis in HIV-infected patients?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	No ART	ART	Relative (95% CI)	Absolute	Quality	Importance
Resolution (important outcome; assessed with clinical exam; follow-up 22 months)
1	Observational study	Very serious¹	No serious inconsistency	Serious²	No serious imprecision	Undetected	7/17 (41.2%)	16/19 (84.2%)	RR 2.05 (1.12 to 3.73)	432 more per 1000 (from 49 more to 1000 more)	⊕○○○ VERY LOW	MODERATE
Incidence assessed with clinical exam (follow-up 22 months)
1	Observational study	Very serious³	No serious inconsistency	No serious indirectness	No serious imprecision	Undetected	17/44 (38.6%)	19/76 (25%)	RR 0.65 (0.38 to 1.11)	135 fewer per 1000 (from 240 fewer to 43 more)	⊕○○○ VERY LOW³	MODERATE

1: Prospective observational study.

2: This study reports resolution of seborrhoeic dermatitis among those on ART who got seborrhoeic dermatitis after a certain follow-up period.

3: No explanation was provided.

Table 2.5Should ART be used for seborrhoeic dermatitis in HIV-infected patients?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	No ART	ART	Relative (95% CI)	Absolute	Quality	Importance
Incidence assessed by clinical examination (follow-up 5 years)
1	Observational study	Very serious¹	No serious inconsistency	Serious²	No serious imprecision	Undetected	—	9/878 (1%)	—	—	⊕○○○ VERY LOW	MODERATE

1: Subset of cohort.

2: No explanation was provided.

Table 2.6Should ART be used for seborrhoeic dermatitis in HIV-infected patients?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	No ART	ART	Relative (95% CI)	Absolute	Quality	Importance
Incidence assessed by clinical examination (follow-up 8 weeks)
1	Observational study	No serious risk of bias	No serious inconsistency	Very serious¹	No serious imprecision	Undetected	15/43 (34.9%)	2/10 (20%)	RR 0.22 (0.05 to 0.89)	272 fewer per 1000 (from 38 fewer to 331 fewer)	⊕○○○ VERY LOW	MODERATE

1: Subgroup analysis of late initiation of ART.

Table 3.1Should ART alone or with other treatments versus no intervention be used in HIV-positive patients with PPE?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	ART alone or with other treatments	No Intervention	Relative (95% CI)	Absolute	Quality	Importance
Reduction in PPE severity (follow-up median 24 months; measured with: defined by sum of day and night itch scores; range of scores: 0-6; better indicated by higher values)
1	Observational study	No serious risk of bias	No serious inconsistency	Serious¹	Serious²	None	39	53³	—	Mean 0.1 higher (0 to 6 higher)⁴	⊕○○○ VERY LOW	IMPORTANT

1: No comparison with no ART.

2: Small population size, no control.

3: Before and after study.

4: No CIs reported, this is the range of the scoring system.

Table 3.2Should oral therapy with pentoxifylline versus no intervention be used in HIV-positive patients with PPE?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Oral therapy with pentoxifylline	No intervention	Relative (95% CI)	Absolute	Quality	Importance
Pruritus score (follow-up 8 weeks; measured with: visual analog scale, investigator global assessment; range of scores: 0-10; better indicated by lower values)
1	Observational study	No serious risk of bias	No serious inconsistency	No serious indirectness¹	Very serious¹	None	11	12²	—	Mean 3.6 higher (0 to 10 higher)³	⊕○○○ VERY LOW	IMPORTANT

1: Low size of population, no comparison group, no control group.

2: Before and after study.

3: No reported CI, this is the range of the scoring system.

Table 3.3Should dapsone versus antihistamines and topical clobetasol be used in HIV-positive patients with PPE?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Dapsone	Antihistamines and topical clobetasol	Relative (95% CI)	Absolute	Quality	Importance
Favourable response (follow-up 14 weeks; better indicated by lower values)
1	Randomized trial	Very serious¹	No serious inconsistency	No serious indirectness	Very serious²	None	10	10	—	MD 0 higher (0 to 0 higher)³	⊕○○○ VERY LOW	IMPORTANT
Remission period (follow-up 14 weeks; better indicated by lower values)
1	Randomized trial	Very serious¹	No serious inconsistency	No serious indirectness	Very serious⁴	None	10	10	—	MD 0 higher (0 to 0 higher)³	⊕○○○ VERY LOW	IMPORTANT

1: There is no mention of randomization or blinding.

2: There is not enough data to support the result on faster response for each group. The study population is small.

3: No quantitative data reported for this outcome.

4: No data on duration of remission or evaluation of remission among the groups.

Table 3.4Should oral promethazine versus 1% hydrocortisone be used in HIV-positive patients with PPE?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Oral promethazine	1% hydrocortisone	Relative (95% CI)	Absolute	Quality	Importance
Reduction in itch score (measured with: subjective itching score; range of scores: 0–9; better indicated by lower values)
1	Observational study	Serious¹	No serious inconsistency	No serious indirectness	Very serious²	None	50	18	—	Mean 3.9 higher (0 to 9 higher)³	⊕○○○ VERY LOW	IMPORTANT
Reduction in clinical severity score (measured with: score system; range of scores: 1–3; better indicated by higher values)
1	Observational study	Serious¹	No serious inconsistency	No serious indirectness	Serious⁴	None	50	18	—	Mean 1.3 higher (1 to 3 higher)³	⊕○○○ VERY LOW	IMPORTANT

1: Did not specify time to follow-up, or patient characteristics such as adults or children, unclear inclusion criteria.

2: There was no reported value for significant difference between scores at the start and end of the treatment. No control group.

3: There was no CI data, this is the range of the score system.

4: No reported data on differences at the start and end of study.

Table 4.1Should ART and isotretinoin versus no intervention be used in HIV-positive patients with eosinophilic folliculitis?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	ART and isotretinoin	Control	Relative (95% CI)	Absolute	Quality	Importance
Resolution of lesions (follow-up 36 months)
1	Observational study¹	Serious²	No serious inconsistency	No serious indirectness	Serious²^,³	None	16/23 (69.6%)	—	—	—	⊕○○○ VERY LOW	IMPORTANT
1	Observational study¹	Serious²	No serious inconsistency	No serious indirectness	Serious²^,³	None		0%	—	—	⊕○○○ VERY LOW	IMPORTANT

1: Prospective study.

2: No controls.

3: No precise definition on complete response and partial response criteria.

*: Only case reports and retrospective data available in the review. The recommendations were made on the basis of expert consensus.

Table 5.1Should terbinafine 1% cream/gel versus placebo cream/gel be used for tinea cruris and tinea corporis?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Terbinafine 1% cream/gel	Placebo cream/gel	Relative (95% CI)	Absolute	Quality	Importance
Mycological cure (assessed with: negative potassium hydroxide [KOH] microscopy, or culture, or both. Treatment duration 1–2 weeks)
7	Randomized trials	Very serious¹	Serious²	No serious indirectness	No serious imprecision	None	151/168 (89.9%)	42/162 (25.9%)	Not pooled	Not pooled	⊕○○○ VERY LOW	CRITICAL
7	Randomized trials	Very serious¹	Serious²	No serious indirectness	No serious imprecision	None		31.3%	Not pooled	Not pooled	⊕○○○ VERY LOW	CRITICAL
Clinical cure (follow-up 2-4 weeks; assessed with: resolution of clinical signs and symptoms. Treatment duration 1–2 weeks)
5	Randomized trials	Serious³	No serious inconsistency	No serious indirectness	Serious⁴	None⁵	104/134 (77.6%)	23/139 (16.5%)	RR 4.51 (3.1 to 6.56)	581 more per 1000 (from 347 more to 920 more)	⊕⊕○○ LOW	CRITICAL
5	Randomized trials	Serious³	No serious inconsistency	No serious indirectness	Serious⁴	None⁵		13.3%	RR 4.51 (3.1 to 6.56)	467 more per 1000 (from 279 more to 739 more)	⊕⊕○○ LOW	CRITICAL
Adverse effects (follow-up 0-8 weeks; assessed with: reported by investigators and/or participants)
7	Randomized trials	Serious¹	No serious inconsistency	No serious indirectness	Very serious⁶	None	8/232 (3.4%)	23/237 (9.7%)	RR 0.43 (0.2 to 0.92)	55 fewer per 1000 (from 8 fewer to 78 fewer)	⊕○○○ VERY LOW	CRITICAL
7	Randomized trials	Serious¹	No serious inconsistency	No serious indirectness	Very serious⁶	None		2.9%	RR 0.43 (0.2 to 0.92)	17 fewer per 1000 (from 2 fewer to 23 fewer)	⊕○○○ VERY LOW	CRITICAL
Relapse or recurrence (follow-up 1–8 weeks; assessed with: evidence of clinical or mycological infection in previously cured participants)
3	Randomized trials	Serious⁷	No serious inconsistency	No serious indirectness	Serious⁸	None	0/81 (0%)	0/87 (0%)	Not pooled	Not pooled	⊕⊕○○ LOW	IMPORTANT
3	Randomized trials	Serious⁷	No serious inconsistency	No serious indirectness	Serious⁸	None		0%	Not pooled	Not pooled	⊕⊕○○ LOW	IMPORTANT
Participant-judged cure (assessed with: judgement of treatment as ‘good’ or ‘very good’)
2	Randomized trials	Serious⁹	No serious inconsistency	No serious indirectness	Serious¹⁰	None¹¹	110/122 (90.2%)	26/131 (19.8%)	RR 4.46 (3.16 to 6.31)	687 more per 1000 (from 429 more to 1000 more)	⊕⊕○○ LOW	IMPORTANT
2	Randomized trials	Serious⁹	No serious inconsistency	No serious indirectness	Serious¹⁰	None¹¹		0%	RR 4.46 (3.16 to 6.31)	—	⊕⊕○○ LOW	IMPORTANT

1: Random sequence generation, allocation concealment and blinding at unclear risk of bias across studies, with 2 studies (Lebwohl et al., 2001; Millikan, 1990) judged overall at high risk of bias. In both of these studies, there was a high dropout rate (20–25%) in already underpowered studies.

2: Substantial unexplained heterogeneity.

3: 3 studies (Lebwohl et al., 2001; Millikan, 1990; Zaias et al., 1993) judged at high risk of bias overall.

4: Small sample size, optimal size would be 2790 participants.

5: Although there is a large effect (RR 4.51, in all studies RR > 4.00), there are threats to validity, see risk of bias.

6: CI includes the threshold for appreciable benefit (0.75) and nearly no effect (1.0), very low number of events, low sample size (optimal size would be 4238 participants).

7: Millikan (1990) judged at high risk of bias overall – high dropout rate in an underpowered study; sequence generation, allocation concealment and blinding judged at unclear risk of bias in remaining studies.

8: Low number of events, sample size is lower than optimal.

9: Blinding for both studies judged at unclear risk of bias, and Zaias and colleagues (1993) judged overall at high risk of bias – details on total number of randomized participants not given.

10: Number of events <300 and optimal size would be 2210 participants.

11: Although there is a large effect (RR > 2), there are threats to validity, see risk of bias.

Table 5.2Should naftifine 1% cream once or twice daily versus placebo cream once or twice daily be used for tinea cruris and tinea corporis?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Naftifine 1% cream once or twice daily	Placebo cream once or twice daily	Relative (95% CI)	Absolute	Quality	Importance
Mycological cure (assessed with: negative KOH microscopy and culture. Treatment duration 2–4 weeks)
3	Randomized trials	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None³	83/95 (87.4%)	33/92 (35.9%)	RR 2.38 (1.8 to 3.14)	495 more per 1000 (from 287 more to 768 more)	⊕⊕○○ LOW	CRITICAL
3	Randomized trials	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None³		32.1%	RR 2.38 (1.8 to 3.14)	443 more per 1000 (from 257 more to 687 more)	⊕⊕○○ LOW	CRITICAL
Clinical cure (follow-up 6 weeks; assessed with: resolution of clinical signs and symptoms at least 2 weeks from start of treatment)
1	Randomized trials	No serious risk of bias	No serious inconsistency	No serious indirectness	Very serious⁴	None⁵	25/32 (78.1%)	10/31 (32.3%)	RR 2.42 (1.41 to 4.16)	458 more per 1000 (from 132 more to 1000 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trials	No serious risk of bias	No serious inconsistency	No serious indirectness	Very serious⁴	None⁵		0%	RR 2.42 (1.41 to 4.16)	—	⊕⊕○○ LOW	CRITICAL
Adverse effects (follow-up 0–6 weeks; assessed with: reported by investigators and/or participants)
3	Randomized trials	Serious¹	No serious inconsistency	No serious indirectness	Veryserious⁶	None	3/99 (3%)	7/96 (7.3%)	RR 0.44 (0.13 to 1.57)	41 fewer per 1000 (from 63 fewer to 42 more)	⊕○○○ VERY LOW	CRITICAL
3	Randomized trials	Serious¹	No serious inconsistency	No serious indirectness	Veryserious⁶	None		5.4%	RR 0.44 (0.13 to 1.57)	30 fewer per 1000 (from 47 fewer to 31 more)	⊕○○○ VERY LOW	CRITICAL
Relapse or recurrence (follow-up 6 weeks; assessed with: evidence of clinical or mycological infection in previously cured participants)
1	Randomized trials	No serious risk of bias	No serious inconsistency	No serious indirectness	Very serious⁷	None	0/30 (0%)	3/14 (21.4%)	RR 0.07 (0 to 1.25)	199 fewer per 1000 (from 214 fewer to 54 more)	⊕⊕○○ LOW	IMPORTANT
								0%		—
								0%		—

1: Dobson and colleagues (1991) judged at high risk of bias overall – high dropout rate (27%) in an already underpowered study. Numbers of participants in each group after randomization unclear.

2: Low sample size, optimal size would be 938.

3: Although large treatment effect (RR > 2), there were threats to validity, see risk of bias.

4: Very low total number of participants, optimal size would be 1114, and wide CI.

5: Although large treatment effect (RR > 2), there were threats to validity, see imprecision.

6: CI includes appreciable harm, no effect and appreciable benefit. Furthermore, low number of events and small sample size (optimal size would be 5804 participants).

7: CI includes appreciable harm, no effect and appreciable benefit. Furthermore, low sample size (optimal size would be 1608 participants).

Table 5.3Should azoles versus moderate-potent corticosteroid/azole combinations be used for tinea cruris and tinea corporis?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Azoles	Moderate-potent corticosteroid/azole combinations	Relative (95% CI)	Absolute	Quality	Importance
Mycological cure (assessed with: negative KOH microscopy and culture. Treatment duration 2–3 weeks)
6¹	Randomized trials	Serious²	No serious inconsistency	Serious³	No serious imprecision	None	245/313 (78.3%)	248/312 (79.5%)	RR 0.99 (0.93 to 1.05)	8 fewer per 1000 (from 56 fewer to 40 more)	⊕⊕○○ LOW	CRITICAL
6¹	Randomized trials	Serious²	No serious inconsistency	Serious³	No serious imprecision	None		88.1%	RR 0.99 (0.93 to 1.05)	9 fewer per 1000 (from 62 fewer to 44 more)	⊕⊕○○ LOW	CRITICAL
Clinical cure (immediately at end of treatment) (assessed with: resolution of clinical signs and symptoms at least 2 weeks from the start of treatment)
4⁴	Randomized trials	Serious²	No serious inconsistency	Serious³	Serious⁵	None	90/181 (49.7%)	133/172 (77.3%)	RR 0.67 (0.53 to 0.84)	255 fewer per 1000 (from 124 fewer to 363 fewer)	⊕○○○ VERY LOW	CRITICAL
4⁴	Randomized trials	Serious²	No serious inconsistency	Serious³	Serious⁵	None		83.6%	RR 0.67 (0.53 to 0.84)	276 fewer per 1000 (from 134 fewer to 393 fewer)	⊕○○○ VERY LOW	CRITICAL
Adverse effects (follow-up 0–4 weeks; assessed with: reported by investigators and/or participants)
5⁶	Randomized trials	Serious²	No serious inconsistency	No serious indirectness	Very serious⁷	None	18/336 (5.4%)	13/332 (3.9%)	RR 1.36 (0.68 to 2.69)	14 more per 1000 (from 13 fewer to 66 more)	⊕○○○ VERY LOW	CRITICAL
								1.8%		6 more per 1000 (from 6 fewer to 30 more)
								0%		—
Participant-judged cure (assessed with: 4 point symptom score scale)
1⁸	Randomized trials	Very serious⁹	No serious inconsistency	No serious indirectness	No serious imprecision	None	—	—	—	—	⊕⊕○○ LOW	IMPORTANT
								0%		—
								0%		—

1: Katz et al., 1984; Li et al., 2004; Pariser & Pariser, 1995; Shen et al., 2002; Wang et al., 2000; Wortzel, 1982.

2: Sequence generation, allocation concealment and blinding at unclear risk of bias across all studies. Pariser & Pariser (1995) judged overall at high risk of bias.

3: 4 different azole creams and 2 different corticosteroid/azole creams assessed in these studies.

4: Pariser & Pariser, 1995; Shen et al., 2002; Wang et al., 2000; Wortzel, 1982.

5: Low sample size, optimal sample size would be 500, CI includes threshold 0.75.

6: Katz et al., 1984; Li et al., 2004; Pariser & Pariser, 1995; Shen et al., 2002; Wortzel, 1982.

7: CI includes appreciable harm, no effect and appreciable benefit. Furthermore, low number of events and very small sample size (optimal sample size would be 10 310).

8: Pariser & Pariser, 1995.

9: Blinding judged at unclear risk of bias, and minimal data were reported on patient-judged cure.

Table 5.4Should azoles versus allylamines be used for tinea cruris and tinea corporis?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Azoles	Allylamines	Relative (95% CI)	Absolute	Quality	Importance
Mycological cure (assessed with: negative KOH microscopy and culture. Treatment duration 1–7 weeks)
7	Randomized trials	Serious¹	Serious²	Serious³	No serious imprecision	None	288/323 (89.2%)	303/315 (96.2%)	Not pooled	Not pooled	⊕○○○ VERY LOW	CRITICAL
7	Randomized trials	Serious¹	Serious²	Serious³	No serious imprecision	None		100%	Not pooled	Not pooled	⊕○○○ VERY LOW	CRITICAL
Clinical cure (assessed with: resolution of clinical signs and symptoms at least 2 weeks from the start of treatment. Treatment duration 1–7 weeks)
6	trials Randomized	Serious⁴	Serious²	Serious³	No serious imprecision	None	249/305 (81.6%)	270/300 (90%)	Not pooled	Not pooled	⊕○○○ VERY LOW	CRITICAL
6	trials Randomized	Serious⁴	Serious²	Serious³	No serious imprecision	None		92.3%	Not pooled	Not pooled	⊕○○○ VERY LOW	CRITICAL
Adverse effects (follow-up 0–8 weeks; assessed with: reported by investigators and/or participants)
5	Randomized trials	Serious¹	No serious inconsistency	No serious indirectness	Very serious⁵	None	2/197 (1%)	4/189 (2.1%)	RR 0.7 (0.18 to 2.68)	6 fewer per 1000 (from 17 fewer to 36 more)	⊕○○○ VERY LOW	CRITICAL
5	Randomized trials	Serious¹	No serious inconsistency	No serious indirectness	Very serious⁵	None		2.2%	RR 0.7 (0.18 to 2.68)	7 fewer per 1000 (from 18 fewer to 37 more)	⊕○○○ VERY LOW	CRITICAL
Relapse or recurrence (follow-up 2–4 weeks; assessed with: evidence of clinical and mycological relapse after the end of treatment. Assessed in 3 studies)
3⁶	Randomized trials	Serious¹	No serious inconsistency	No serious indirectness	Very serious⁷	None	1/61 (1.6%)	0%	RR 2.33 (0.21 to 26.23)	—	⊕○○○ VERY LOWW	IMPORTANTT
Duration of treatment until clinical cure (range of scores: 21–77; better indicated by lower values)
1	Randomized trial	Serious⁸	No serious inconsistency	No serious indirectness	Very serious⁹	None	2	5	—	MD 33.60 lower (46.91 too 20.29 lower)	⊕○○○ VERY LOW	IMPORTANT

1: Haroon and colleagues (1996) and Jerajani and colleagues (2013) were both open trials, and blinding was therefore judged at high risk of bias. In addition, the attrition rate was also high in both studies (20% and 25% respectively). Sequence generation, allocation concealment and blinding all judged at an unclear risk of bias in remaining studies.

2: Substantial heterogeneity (I² = 75%).

3: Six different azole creams used across the studies. Allylamine treatment regimens different across all studies.

4: Jerajani and colleagues (2013) judged at high risk of bias due to lack of blinding and high attrition rate. Sequence generation, allocation concealment and blinding for the remaining studies all judged at an unclear risk of bias.

5: Low number of events, CI is wide, including appreciable harm, no effect and appreciable benefit, optimal size would be 15 414 participants.

6: Hantschke & Reichenberger (1980); Haroon etal., 1996; Jerajani et al., 2013.

7: Low total number of participants and wide CI including no effect and appreciable harm.

8: Sequence generation, allocation concealment and blinding all judged at unclear risk of bias.

9: Only 9 participants in total.

Table 5.5Should azoles versus benzylamines be used for tinea cruris and tinea corporis?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Azoles	Benzylamines	Relative (95% CI)	Absolute	Quality	Importance
Mycological cure (assessed with: negative KOH microscopy and culture. Treatment duration 2–4 weeks)
3¹	Randomized trials	Serious²	No serious inconsistency	Serious³	No serious imprecision	None	107/112 (95.5%)	101/107 (94.4%)	RR 1.01 (0.94 to 1.07)	9 more per 1000 (from 57 fewer to 66 more)	⊕⊕○○ LOW	CRITICAL
3¹	Randomized trials	Serious²	No serious inconsistency	Serious³	No serious imprecision	None		931%	RR 1.01 (0.94 to 1.07)	9 more per 1000 (from 56 fewer to 65 more)	⊕⊕○○ LOW	CRITICAL
Clinical cure (assessed with: resolution of clinical signs and symptoms at least 2 weeks from the start of treatment)
2⁴	Randomized trials	No serious risk of bias	Serious⁵	No serious indirectness	Serious⁶	None	47/84 (56%)	45/85 (52.9%)	Not pooled	Not pooled	⊕⊕○○ LOW	CRITICAL
2⁴	Randomized trials	No serious risk of bias	Serious⁵	No serious indirectness	Serious⁶	None		0%	Not pooled	Not pooled	⊕⊕○○ LOW	CRITICAL
Adverse effects (follow-up 0–8 weeks; assessed with: reported by investigators and/or participants)
3¹	Randomized trials	Serious²	No serious inconsistency	No serious indirectness	Very serious⁷	None	12/131 (9.2%)	14/132 (10.6%)	RR 0.85 (0.41 to 1.76)	16 fewer per 1000 (from 63 fewer to 81 more)	⊕○○○ VERY LOW	CRITICAL
3¹	Randomized trials	Serious²	No serious inconsistency	No serious indirectness	Very serious⁷	None		10.3%	RR 0.85 (0.41 to 1.76)	15 fewer per 1000 (from 61 fewer to 78 more)	⊕○○○ VERY LOW	CRITICAL
Relapse or recurrence (follow-up 4–8 weeks; assessed with: evidence of clinical or mycological disease after successful treatment)
3¹	Randomized trials	Serious²	No serious inconsistency	Serious³	Serious⁸	None	4/110 (3.6%)	2/105 (1.9%)	RR 1.84 (0.35 to 9.6)	16 more per 1000 (from 12 fewer to 164 more)	⊕○○○ VERY LOW	IMPORTANT
								0%		—
								0%		—

1: Ramam et al., 2003; Singal et al., 2005; Li et al., 2006.

2: Ramam and colleagues (2003) judged at high risk of bias overall – high attrition rate and study funded by industry supplying both interventions.

3: Different azoles assessed in the studies.

4: Singal et al., 2005; Li et al., 2006.

5: Substantial heterogeneity apparent.

6: Low total number of participants.

7: Low sample size, optimal size is around 5500, and CI includes both no effect and appreciable harm.

8: Very wide CI, low event rate, small sample size.

Table 5.6Should azoles versus placebo be used for tinea cruris and tinea corporis?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Azoles	Placebo	Relative (95% CI)	Absolute	Quality	Importance
Mycological cure (assessed with: negative KOH microscopy and culture. Treatment duration 2-4 weeks)
4¹^,²	trials Randomized	Serious³	Serious⁴	Serious⁵	No serious imprecision	None	257/284 (90.5%)	72/206 (35%)	Not pooled	Not pooled	⊕○○○ VERY LOW	CRITICAL
4¹^,²	trials Randomized	Serious³	Serious⁴	Serious⁵	No serious imprecision	None		0%	Not pooled	Not pooled	⊕○○○ VERY LOW	CRITICAL
Clinical cure (assessed with: resolution of clinical signs and symptoms. Treatment duration 2-4 weeks)
3²^,⁶	Randomized trials	Serious⁷	Serious⁴	Serious⁵	No serious imprecision	None	153/211 (72.5%)	34/125 (27.2%)	Not pooled	Not pooled	⊕○○○ VERY LOW	CRITICAL
3²^,⁶	Randomized trials	Serious⁷	Serious⁴	Serious⁵	No serious imprecision	None		0%	Not pooled	Not pooled	⊕○○○ VERY LOW	CRITICAL
Adverse effects (follow-up 0-5 weeks; assessed with: reported by investigators and/or participants)
3²^,⁶	Randomized trials	Serious⁷	No serious inconsistency	No serious indirectness	Very serious⁸	None	2/132 (1.5%)	11/134 (8.2%)	RR 0.25 (0.06 to 0.99)	62 fewer per 1000 (from 1 fewer to 77 fewer)	⊕○○○ VERY LOW	CRITICAL
								0%		—
								0%		—

1: Bagatell, 1986; Miura et al., 1979; Spiekermann & Young, 1976; Tanenbaum et al., 1989.

2: Miura and colleagues (1979) - 2 comparisons (econazole versus placebo and clotrimazole versus placebo).

3: Spiekernann &Young (1976) judged at high risk of bias overall due to high attrition rate (33%) and industry funded study. Sequence generation, allocation concealment and blinding all judged at unclear risk of bias across remaining studies.

4: Substantial unexplained heterogeneity, data not pooled.

5: 4 different azoles.

6: Bagatell, 1986; Miura et al.,1979;Tanenbaum et al., 1989.

7: Sequence generation, allocation concealment and blinding all judged at unclear risk of bias across all studies.

8: Wide CI including appreciable harm and low sample size, optimal size around 4724.

Table 6.1Should acyclovir versus placebo be used for herpes zoster in HIV-infected adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	No acyclovir	Acyclovir	Relative (95% CI)	Absolute	Quality	Importance
Cessation of new lesions
5	Randomized trials	Serious risk of bias¹	No serious inconsistency	Serious indirectness²	No serious imprecision	Undetected	122/200 (61%)	127/206 (61.7%)	HR 1.54 (1.21 to 1.97)	155 more per 1000 (from 70 more to 234 more)	⊕⊕○○ LOW	CRITICAL
Mean time new lesion stoppage (better indicated by lower values)
2	Randomized trials	Serious risk of bias¹	No serious inconsistency	Serious indirectness²	No serious imprecision	Undetected	53	55	—	Mean time to new lesion stoppage in intervention group was 0.59 lower (1.11 to 0.08 lower)	⊕⊕○○ LOW	CRITICAL
Lesion healing (critical outcome)
3	Randomized trials	Serious risk of bias¹	No serious inconsistency	Serious indirectness²	No serious imprecision	Undetected	78/148 (52.7%)	66/135 (48.9%)	HR 1.48 (1.06 to 2.05)	143 more per 1000 (from 21 more to 257 more)	⊕⊕○○ LOW	CRITICAL
Mean time to full crusting (better indicated by lower values)
1	Randomized trial	Serious risk of bias¹	No serious inconsistency	Serious indirectness²	No serious imprecision	Undetected	29	29		Mean time to full crusting in intervention group was 7.4 lower (15.78 lower to 0.98 higher)	⊕⊕○○ LOW	CRITICAL

1: Details of randomization not provided or incomplete outcome data.

2: Studies involved non-HIV populations.

Table 6.2Should famciclovir versus acyclovir be used for herpes zoster in HIV-infected adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	Acyclovir	Famciclovir	Relative (95% CI)	Absolute	Quality	Importance
Lesion healing (critical outcome)
3	Randomized trials	Serious risk of bias¹	No serious inconsistency	Serious indirectness²	No serious imprecision	Undetected	291/294 (99%)	271/275 (98.5%)	HR 1.18 (0.98 to 1.41)	6 more per 1000 (from 1 fewer to 9 more)	⊕⊕○○ LOW	CRITICAL

1: Details of randomization not provided.

2: Studies involved non-HIV populations.

Table 6.3Should famciclovir versus acyclovir be used for stoppage of new herpes zoster lesions in HIV-infected adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	No acyclovir	Acyclovir	Relative (95% CI)	Absolute	Quality	Importance
Cessation of new lesions (critical outcome)
1	Randomized trial	Serious risk of bias¹	No serious inconsistency	Serious indirectness²	No serious imprecision	Undetected	77/128 (60.2%)	69/121 (57%)	HR 1.45 (0.93 to 2.26)	135 more per 1000 (from 27 fewer to 273 more)	⊕⊕○○ LOW	CRITICAL

1: Details of randomization not provided.

2: Studies involved non-HIV populations.

Table 6.4Should brivudin versus acyclovir be used for herpes zoster in HIV-infected adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	Acyclovir	Brivudin	Relative (95% CI)	Absolute	Quality	Importance
Cessation of new lesions (critical outcome)
2	Randomized trials	Serious risk of bias¹	No serious inconsistency	Serious indirectness²	No serious imprecision	Undetected	633/633 (100%)	635/635 (100%)	HR 1.11 (0.99 to 1.24)	0 fewer per 1000	⊕⊕○○ LOW	CRITICAL
Lesion healing (critical outcome)
3	Randomized trials	Serious risk of bias¹	No serious inconsistency	Serious indirectness²	No serious imprecision	Undetected	653/657 (99.4%)	653/658 (99.2%)	HR 0.95 (0.85 to 1.06)	2 fewer per 1000 (from 7 fewer to 2 more)	⊕⊕○○ LOW	CRITICAL

1: Details of randomization not provided.

2: Studies involved non-HIV populations.

Table 6.5Should valacyclovir versus acyclovir be used for herpes zoster in HIV-infected adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	Acyclovir	Brivudin	Relative (95% CI)	Absolute	Quality	Importance
Lesion healing (critical outcome)
2	Randomized trials	Serious risk of bias¹	No serious inconsistency	Serious indirectness²	No serious imprecision	Undetected	391/406 (96.3%)	403/414 (97.3%)	HR 1.01 (0.88 to 1.16)	1 more per 1000 (from 18 fewer to 15 more)	⊕⊕○○ LOW	CRITICAL
Cessation of new lesions (critical outcome)
1	Randomized trial	Serious risk of bias¹	No serious inconsistency	Serious indirectness²	No serious imprecision	Undetected	376/376 (100%)	384/384 (100%)	HR 1.03 (0.89 to 1.2)	1000 per 1000	⊕⊕○○ LOW	CRITICAL

1: Details of randomization not provided or incomplete outcome data.

2: Studies involved non-HIV populations.

Table 7.1Should permethrin (topical) versus ivermectin (oral 1 dose) be used for scabies?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Permethrin topical	Ivermectin oral 1 dose	Relative (95% CI)	Absolute	Quality	Importance
Complete cure (follow-up 4 weeks; assessed with: reduction in both the number of lesions as well as the grade of pruritus by more than or equal to 50% )
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Very serious¹	None	36/38 (94.7%)	36/40 (90%)	RR 1.06 (0.95 to 1.67)	54 more per 1000 (from 45 fewer to 603 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Very serious¹	None		0%	RR 1.06 (0.95 to 1.67)	—	⊕⊕○○ LOW	CRITICAL

1: Poor confirmation of diagnosis, unclear definition of complete clinical cure, and rate of cure defined as >50% improvement in lesion count.

Table 7.2Should oral ivermectin with antihistaminics versus permethrin with antihistaminics be used for scabies?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Oral ivermectin with antihistaminics	Permethrin with antihistaminics	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure rate (follow-up 4 weeks; measured with: number of lesions;¹better indicated by lower values)
1	Randomized trial	Serious¹^,²	No serious inconsistency	No serious indirectness	Very serious¹^,²	None	100	103	—	RR 1.01 higher (0.95 to 1.08 higher)	⊕○○○ VERY LOW	CRITICAL

1: Rate of clinical cure not clearly defined.

2: Mention of a single dose in methodology and repeated dose in discussion for oral ivermectin.

Table 7.3Should oral ivermectin (two applications) versus topical permethrin (two applications) be used for scabies?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Oral ivermectin two applications	Topical permethrin two applications	Relative (95% CI)	Absolute	Quality	Importance
Cure (follow-up 4 weeks; assessed with: disappearance of itching, clearance of skin lesions and absence of mites on microscopy skin lesions)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Very serious²	None	44/50 (88%)	42/50 (84%)	RR 1.06 (0.92 to 1.33)	50 more per 1000 (from 67 fewer to 277 more)	⊕○○○ VERY LOW	CRITICAL
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Very serious²	None		0%	RR 1.06 (0.92 to 1.33)	—	⊕○○○ VERY LOW	CRITICAL

1: Not blinded.

2: Unclear definition of number of lesions considered as cure. Second dose of each treatment only provided to non-responsive patients.

Table 7.4Should topical permethrin (two applications) versus oral ivermectin (two applications) be used for scabies?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Topical permethrin two applications	Oral ivermectin two applications	Relative (95% CI)	Absolute	Quality	Importance
Cure rate (follow-up 2 weeks; measured with: lesion count;¹ better indicated by lower values)
1	Randomized trial	Serious²	No serious inconsistency	No serious indirectness	Very serious¹^,³^,⁴	None	28	27	—	Rate % 96.43 higher (0 to 100 higher)⁵	⊕○○○ VERY LOW	CRITICAL

1: Not clearly defined, “The participants who did not have any new lesions were considered as cured”.

2: Not blinded.

3: Does not state how many patients received treatment with azithromycin.

4: Only patients who did not improve at week 1 received second application.

5: The rate of cure can go from 0% to 100%.

Table 7.5Should topical permethrin (two applications) versus benzyl benzoate (two applications) be used for scabies?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Topical permethrin two applications	Benzyl benzoate two applications	Relative (95% CI)	Absolute	Quality	Importance
Cure rate (follow-up 2 weeks; measured with: lesion count;¹ better indicated by lower values)
1	Randomized trial	Serious²	No serious inconsistency	No serious indirectness	Very serious³^,⁴^,⁵	None	34	35	—	Rate % 96.46 higher (0 to 100 higher)⁶	⊕○○○ VERY LOW	CRITICAL

1: Not clearly defined, “The participants who did not have any new lesions were considered as cured”.

2: Not blinded.

3: Not clearly defined, “The participants who did not have any new lesions were considered as cured”.

4: Does not state how many patients received treatment with azithromycin.

5: Only patients who did not improve at week 1 received second application.

6: The rate of cure can go from 0% to 100%.

Table 7.6Should oral ivermectin (two applications) versus benzyl benzoate (two applications) be used for scabies?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Oral ivermectin two applications	Benzyl benzoate two applications	Relative (95% CI)	Absolute	Quality	Importance
Cure rate (follow-up 2 weeks; measured with: lesion count;¹ better indicated by lower values)
1	Randomized trial	Serious²	No serious inconsistency	No serious indirectness	Very serious¹^,³^,⁴	None	34	35	—	Rate % 100 higher (0 to 100 higher)	⊕○○○ VERY LOW	CRITICAL

1: Not clearly defined, “The participants who did not have any new lesions were considered as cured”.

2: Not blinded.

3: Does not state how many patients received treatment with azithromycin.

4: Only patients who did not improve at week 1 received second application.

Table 7.7Should ivermectin be used in HIV-positive patients with scabies?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Ivermectin	Control	Relative (95% CI)	Absolute	Quality	Importance
Cure (follow-up 4 months; assessed with: patients with no pruritus, no dermatological evidence of scabies and no positive scrapings)
1	Observational study	Serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	None	11/11 (100%)	11/13 (84.6%)	RR 3 (0.13 to 66.5)	1000 more per 1000 (from 736 fewer to 1000 more)²	⊕○○○ VERY LOW	CRITICAL
1	Observational study	Serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	None		0%	RR 3 (0.13 to 66.5)	—	⊕○○○ VERY LOW	CRITICAL

1: Not randomized, small study population.

Table 7.8Should oral ivermectin versus benzyl benzoate solution be used in HIV-positive patients with scabies?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Oral ivermectin	Benzyl benzoate solution	Relative (95% CI)	Absolute	Quality	Importance
Complete clinical response (follow-up 4 weeks; assessed with: resolution of itching and dermatological or microbiological cure)
1	Observational study¹	Serious²	No serious inconsistency	No serious indirectness	Serious³	None	12/21 (57.1%)	10/22 (45.5%)	RR 0.78 (0.42 to 1.46)	100 fewer per 1000 (from 264 fewer to 209 more)	⊕○○○ VERY LOW	CRITICAL
1	Observational study¹	Serious²	No serious inconsistency	No serious indirectness	Serious³	None		0%	RR 0.78 (0.42 to 1.46)	—	⊕○○○ VERY LOW	CRITICAL

1: Observational.

2: Retrospective study, small study population.

3: No clear definition of dermatological cure in respect to the number of lesions.

Table 7.9Should oral ivermectin versus a combination of topical benzyl benzoate and oral ivermectin be used in HIV-positive patients with scabies?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Oral ivermectin	Combination of topical benzyl benzoate and oral ivermectin	Relative (95% CI)	Absolute	Quality	Importance
Complete clinical response (follow-up 4 weeks; assessed with: resolution of itching and dermatological or microbiological cure)
1	Observational study¹	Serious¹	No serious inconsistency	No serious indirectness	Serious¹^,²	None	12/21 (57.1%)	17/17 (100%)	RR 0.06 (0.004 to 1.03)	940 fewer per 1000 (from 996 fewer to 30 more)	⊕○○○ VERY LOW	CRITICAL
1	Observational study¹	Serious¹	No serious inconsistency	No serious indirectness	Serious¹^,²	None		0%	RR 0.06 (0.004 to 1.03)	—	⊕○○○ VERY LOW	CRITICAL

1: Retrospective study.

2: Small study population.

Table 7.10Should ivermectin versus combined therapy (ivermectin, permethrin, salicylic acid) be used for crusted scabies?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Ivermectin	Combined therapy of ivermectin, permethrin, salicylic acid	Relative (95% CI)	Absolute	Quality	Importance
Cure (assessed with: elimination of lesions)
1	Observational study¹	Very serious²	No serious inconsistency	No serious indirectness	Serious³	None	6/8 (75%)	2/2 (100%)	RR 1.66 (0.10 to 25.8)	660 more per 1000 (from 900 fewer to 1000 more)	⊕○○○ VERY LOW
1	Observational study¹	Very serious²	No serious inconsistency	No serious indirectness	Serious³	None		0%	RR 1.66 (0.10 to 25.8)	—	⊕○○○ VERY LOW

1: Retrospective.

2: Small study population.

3: Repeated doses were administered to unresponsive patients. The combination group received 3 doses of ivermectin and only received combined therapy when single doses of ivermectin did not resolve completely.

Table 7.11Should ivermectin (single dose) versus benzyl benzoate (two applications) be used for scabies?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Ivermectin single dose	Benzyl benzoate two applications	Relative (95% CI)	Absolute	Quality	Importance
Cure (follow-up median 28 days; assessed with: not clearly defined)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	16/65 (24.6%)	37/68 (54.4%)	OR 0.23 (0.10 to 0.50)	329 fewer per 1000 (from 170 fewer to 437 fewer)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None		0%	OR 0.23 (0.10 to 0.50)	—	⊕⊕○○ LOW	CRITICAL

1: Not blinded.

2: No HIV test was performed before or after, or immunological essay.

Table 8.1Should cryotheraphy versus cryotheraphy and podophyllotoxin be used in HIV-positive patients with molluscum contagiosum?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Cryotheraphy	Cryotheraphy and podophyllotoxin	Relative (95% CI)	Absolute	Quality	Importance
Lesion elimination (follow-up 1 month; measured with: does not mention method of assessment; better indicated by lower values)
1	Observational study	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	21	19	—	P 0.136 higher (0 to 0 higher)	⊕○○○ VERY LOW	IMPORTANT

1: Does not mention randomization.

2: Does not mention status of HIV infection (e.g. CD4 cell count, viral load).

Table 8.2Should tricholoacetic acid versus cryotherapy be used in HIV-positive patients with molluscum contagiosun?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Tricholoacetic acid	Cryotherapy	Relative (95% CI)	Absolute	Quality	Importance
Reduction in number of lesions (follow-up 8 weeks; measured with: method of evaluation of reduction not specified; better indicated by lower values)
1	Randomized trial¹	Very serious²	No serious inconsistency	No serious indirectness	Serious²	None	20	20	—	Median (%) 90 higher (0 to 0 higher)³	⊕○○○ VERY LOW	IMPORTANT

1: No randomization, each patient received both treatments, one on each side of the face.

2: Not clear if patients were on ART or not.

3: P≤.05 reported but CI was not reported.

Table 9.1Should fluconazole versus ketoconazole be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Fluconazole	Ketoconazole	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure (follow-up mean 1 month)
2	Randomized trials	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	38/42 (90.5%)	29/41 (70.7%)	RR 1.27 (0.97 to 1.66)	191 more per 1000 (from 21 fewer to 467 more)	⊕⊕○○ LOW	CRITICAL
Clinical cure – adults (follow-up mean 1 month)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	17/18 (94.4%)	12/19 (63.2%)	RR 1.5 (1.04 to 2.15)	316 more per 1000 (from 25 more to 726 more)	⊕⊕○○ LOW	IMPORTANT
Clinical cure – children (follow-up mean 4 weeks)
1	Randomized trial	Serious	No serious inconsistency	No serious indirectness	No serious imprecision	None	21/24 (87.5%)	17/22 (77.3%)	RR 1.13 (0.86to1.49)	100 more per 1000 (from 108 fewer to 379 more)	⊕⊕⊕○ MODERATE	CRITICAL
Mycological cure (follow-up mean 1 month)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	13/18 (72.2%)	9/19 (47.4%)	RR 1.52 (0.88 to 2.65)	246 more per 1000 (from 57 fewer to 782 more)	⊕⊕○○ LOW	CRITICAL
Clinical + mycological cure (follow-up mean 1 month)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	17/24 (70.8%)	12/22 (54.5%)	RR 1.3 (0.82 to 2.06)	164 more per 1000 (from 98 fewer to 578 more)	⊕⊕○○ LOW	IMPORTANT

1: Method of randomization not described. Baseline imbalance.

2: Wide CI (includes null and appreciable benefits).

Table 9.2Should fluconazole versus itraconazole be used for the management of HIV-infected adults and children with oral candidiasis?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Fluconazole	Itraconazole	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure: with De Wit et al., 1998 (follow-up mean 1 month)
5	Randomized trials	Very serious¹	Serious²	No serious indirectness	Serious³	None	133/168 (79.2%)	218/306 (71.2%)	RR 1.12 (0.92 to 1.36)	85 more per 1000 (from 57 fewer to 256 more)	⊕○○○ VERY LOW	CRITICAL
5	Randomized trials	Very serious¹	Serious²	No serious indirectness	Serious³	None		68.4%	RR 1.12 (0.92 to 1.36)	82 more per 1000 (from 55 fewer to 246 more)	⊕○○○ VERY LOW	CRITICAL
Clinical cure: without De Wit et al., 1998 (follow-up mean 1 month)
4	Randomized trials	Very serious¹	No serious inconsistency	No serious indirectness	Serious⁴	None	118/148 (79.7%)	214/286 (74.8%)	RR 1.05 (0.94 to 1.16)	37 more per 1000 (from 45 fewer to 120 more)	⊕○○○ VERY LOW	CRITICAL
4	Randomized trials	Very serious¹	No serious inconsistency	No serious indirectness	Serious⁴	None		74.8%	RR 1.05 (0.94 to 1.16)	37 more per 1000 (from 45 fewer to 120 more)	⊕○○○ VERY LOW	CRITICAL
Mycological cure (follow-up mean 1 month)
5	Randomized trials	Serious³	No serious inconsistency	No serious indirectness	Very serious⁵	None	85/168 (50.6%)	145/306 (47.4%)	RR 1.14 (0.9 to 1.46)	66 more per 1000 (from 47 fewer to 218 more)	⊕○○○ VERY LOW	CRITICAL
5	Randomized trials	Serious³	No serious inconsistency	No serious indirectness	Very serious⁵	None		43%	RR 1.14 (0.9 to 1.46)	60 more per 1000 (from 43 fewer to 198 more)	⊕○○○ VERY LOW	CRITICAL
Relapse (follow-up mean 1 month)
5	Randomized trials	Serious⁵	Serious⁵	No serious indirectness	No serious imprecision	None	50/126 (39.7%)	90/207 (43.5%)	RR 0.92 (0.71 to 1.21)	35 fewer per 1000 (from 126 fewer to 91 more)	⊕⊕○○ LOW	CRITICAL
5	Randomized trials	Serious⁵	Serious⁵	No serious indirectness	No serious imprecision	None		44.2%	RR 0.92 (0.71 to 1.21)	35 fewer per 1000 (from 128 fewer to 93 more)	⊕⊕○○ LOW	CRITICAL

1: Lack of blinding of outcome assessors. Method of allocation concealment not reported in De Wit and colleagues (1998). No blinding in De Wit and colleagues (1998).

2: Heterogeneity I2 = 68%; CI includes null.

3: Total number of events is less than 300. 95% CI is wide and includes null.

4: 95% CI includes null.

5: No explanation was provided.

Table 9.3Should fluconazole versus clotrimazole be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Fluconazole	Clotrimazole	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure (follow-up mean 14 days)
2	Randomized trials	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	151/189 (79.9%)	124/169 (73.4%)	RR 1.13 (0.92 to 1.37)	95 more per 1000 (from 59 fewer to 271 more)	⊕⊕○○ LOW	CRITICAL
2	Randomized trials	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None		73.1%	RR 1.13 (0.92 to 1.37)	95 more per 1000 (from 58 fewer to 270 more)	⊕⊕○○ LOW	CRITICAL
Mycological cure (follow-up mean 14 days)
2	Randomized trials	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	100/189 (52.9%)	61/169 (36.1%)	RR 1.47 (1.16 to 1.87)	170 more per 1000 (from 58 more to 314 more)	⊕⊕○○ LOW	CRITICAL
2	Randomized trials	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None		40.5%	RR 1.47 (1.16 to 1.87)	190 more per 1000 (from 65 more to 352 more)	⊕⊕○○ LOW	CRITICAL
Relapse (follow-up mean 14 days)
2	Randomized trials	Serious¹	No serious inconsistency	No serious indirectness	Serious³	None	24/143 (16.8%)	50/107 (46.7%)	RR 0.36 (0.24 to 0.54)	299 fewer per 1000 (from 215 fewer to 355 fewer)	⊕⊕○○ LOW	CRITICAL
2	Randomized trials	Serious¹	No serious inconsistency	No serious indirectness	Serious³	None		34.1%	RR 0.36 (0.24 to 0.54)	218 fewer per 1000 (from 157 fewer to 259 fewer)	⊕⊕○○ LOW	CRITICAL

1: Lack of blinding.

2: Total number of events less than 300.

3: No explanation was provided.

Table 9.4Should fluconazole versus fluconazole stat be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Fluconazole	Fluconazole stat	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure – fluconazole 50 mg daily for 7 days vs fluconazole150 mg stat (follow-up mean 2 weeks)
1	Randomized trial	Very serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	None	26/28 (92.9%)	21/28 (75%)	RR 1.24 (0.98 to 1.57)	180 more per 1000 (from 15 fewer to 428 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Very serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	None		75%	RR 1.24 (0.98 to 1.57)	180 more per 1000 (from 15 fewer to 428 more)	⊕⊕○○ LOW	CRITICAL
Clinical cure – fluconazole 150 mg daily for 14 days vs fluconazole 750 mg stat (follow-up mean 42 days)
1	Randomized trial	Very serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	None	105/110 (95.5%)	105/110 (95.5%)	RR 1 (0.94 to 1.06)	0 fewer per 1000 (from 57 fewer to 57 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Very serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	None		95.5%	RR 1 (0.94 to 1.06)	0 fewer per 1000 (from 57 fewer to 57 more)	⊕⊕○○ LOW	CRITICAL
Mycological cure – fluconazole 50 mg daily for 7 days vs fluconazole 150 mg stat (follow-up mean 2 weeks)
1	Randomized trial	Very serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	None	13/28 (46.4%)	6/28 (21.4%)	RR 2.17 (0.96 to 4.89)	251 more per 1000 (from 9 fewer to 834 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Very serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	None		21.4%	RR 2.17 (0.96 to 4.89)	250 more per 1000 (from 9 fewer to 832 more)	⊕⊕○○ LOW	CRITICAL
Mycological cure – fluconazole 150 mg daily for 14 days vs fluconazole 750 mg stat (follow-up mean 42 days)
1	Randomized trial	Very serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	None	83/110 (75.5%)	93/110 (84.5%)	RR 0.89 (0.78 to 1.02)	93 fewer per 1000 (from 186 fewer to 17 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Very serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	None		84.6%	RR 0.89 (0.78 to 1.02)	93 fewer per 1000 (from 186 fewer to 17 more)	⊕⊕○○ LOW	CRITICAL

1: Allocation concealment not reported; no blinding; random sequence generation not specified.

Table 9.5Should fluconazole versus nystatin be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Fluconazole	Nystatin	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure (follow-up mean 4 days)
1	Randomized trial	Very serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	None	60/83 (72.3%)	36/84 (42.9%)	RR 1.69 (1.27 to 2.23)	296 more per 1000 (from 116 more to 527 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Very serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	None		42.9%	RR 1.69 (1.27 to 2.23)	296 more per 1000 (from 116 more to 528 more)	⊕⊕○○ LOW	CRITICAL
Mycological cure (follow-up mean 48 days)
1	Randomized trial	Very serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	None	41/83 (49.4%)	4/84 (4.8%)	RR 10.37 (3.89 to 27.66)	446 more per 1000 (from 138 more to 1000 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Very serious¹	No serious inconsistency	No serious indirectness	No serious imprecision	None		4.8%	RR 10.37 (3.89 to 27.66)	450 more per 1000 (from 139 more to 1000 more)	⊕⊕○○ LOW	CRITICAL

1: Random sequence generation not reported; allocation concealment not reported; single blind – clinical evaluator at trial sites.

Table 9.6Should D0870 25 mg versus D0870 10 mg be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	D0870: 25 mg	D0870: 10 mg	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure (follow-up mean 2 weeks)
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious imprecision¹	None	2/13 (15.4%)	3/14 (21.4%)	RR 0.97 (0.59 to 1.58)	6 fewer per 1000 (from 88 fewer to 124 more)	⊕⊕⊕○ MODERATE	CRITICAL
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious imprecision¹	None		71.4%	RR 0.97 (0.59 to 1.58)	21 fewer per 1000 (from 293 fewer to 414 more)	⊕⊕⊕○ MODERATE	CRITICAL
Relapse (follow-up mean 2 weeks)
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious imprecision²	None	2/13 (15.4%)	3/14 (21.4%)	RR 0.72 (0.14 to 3.64)	60 fewer per 1000 (from 184 fewer to 566 more)	⊕⊕⊕○ MODERATE	IMPORTANT
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious imprecision²	None		0%	RR 0.72 (0.14 to 3.64)	—	⊕⊕⊕○ MODERATE	IMPORTANT

1: Small sample size.

2: 95% CI includes null. Single study.

Table 9.7Should itraconazole versus clotrimazole be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Itraconazole	Clotrimazole	Relative (95% CI)	Absolute	Quality	Importance
Clinical and mycological cure (follow-up 43 days)
2	Randomized trials	Serious¹	Serious²	No serious indirectness	Serious²	None	43/75 (57.3%)	28/77 (36.4%)	RR 1.34 (0.56 to 3.2)	124 more per 1000 (from 160 fewer to 800 more)	⊕○○○ VERY LOW	CRITICAL
2	Randomized trials	Serious¹	Serious²	No serious indirectness	Serious²	None		50.4%	RR 1.34 (0.56 to 3.2)	171 more per 1000 (from 222 fewer to 1000 more)	⊕○○○ VERY LOW	CRITICAL
Mycological cure (follow-up 43 days)
1	Randomized trial	Serious	Serious	No serious indirectness	Serious	None	39/61 (63.9%)	18/62 (29%)	RR 2.2 (1.43 to 3.39)	348 more per 1000 (from 125 more to 694 more)	⊕○○○ VERY LOW	CRITICAL
1	Randomized trial	Serious	Serious	No serious indirectness	Serious	None		29%	RR 2.2 (1.43 to 3.39)	348 more per 1000 (from 125 more to 693 more)	⊕○○○ VERY LOW	CRITICAL

1: Linpiyawan and colleagues (2000): allocation concealment not reported, single blinded, sequence generation not reported.

2: 95% CI includes 1. I2 = 85%, p < 0.05.

Table 9.8Should melaleuca alcohol-free oral solution versus alcohol-based oral solution be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Melaleuca oral solution: alcohol-free	Alcohol-based	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure (follow-up mean 4 weeks)
1	Randomized trial	Serious	No serious inconsistency	No serious indirectness	Serious¹	None	3/14 (21.4%)	0/13 (0%)	RR 6.53 (0.37 to 115.49)	—	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious	No serious inconsistency	No serious indirectness	Serious¹	None		0%	RR 6.53 (0.37 to 115.49)	—	⊕⊕○○ LOW	CRITICAL
Mycological cure (follow-up mean 4 weeks)
1	Randomized trial	Serious²	No serious inconsistency	No serious indirectness	Serious³	None	9/14 (64.3%)	5/13 (38.5%)	RR 1.67 (0.76 to 3.69)	258 more per 1000 (from 92 fewer to 1000 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious²	No serious inconsistency	No serious indirectness	Serious³	None		38.5%	RR 1.67 (0.76 to 3.69)	258 more per 1000 (from 92 fewer to 1000 more)	⊕⊕○○ LOW	CRITICAL

1: Small sample size, event rate low and 95% CI includes null.

2: Random sequence generation not reported, no blinding. Allocation concealment not reported.

3: No explanation was provided.

Table 9.9Should amphotericin fat emulsion versus glucose solution be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Amphotericin fat emulsion	Glucose solution	Relative (95% CI)	Absolute	Quality	Importance
Clinical score reduction (follow-up mean 4 days; better indicated by lower values)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	11	11	—	Mean difference (MD) 1.1 lower (5.72 lower to 3.52 higher)	⊕⊕○○ LOW	CRITICAL
Mycological cure (follow-up mean 4 days)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious	None	2/11 (18.2%)	2/11 (18.2%)	RR 1 (0.17 to 5.89)	0 fewer per 1000 (from 151 fewer to 889 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious	None		18.2%	RR 1 (0.17 to 5.89)	0 fewer per 1000 (from 151 fewer to 890 more)	⊕⊕○○ LOW	CRITICAL

1: No blinding was used. Unclear sequence generation and allocation concealment.

2: Small study (95% CI wide including null). Small effect size.

Table 9.10Should ketoconazole versus miconazole be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Ketoconazole	Miconazole	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure (follow-up mean 14 days)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	159/179 (88.8%)	155/178 (87.1%)	RR 1.02 (0.94 to 1.1)	17 more per 1000 (from 52 fewer to 87 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None		0%	RR 1.02 (0.94 to 1.1)	—	⊕⊕○○ LOW	CRITICAL
Relapse (follow-up mean 14 days)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	34/148 (23%)	45/146 (30.8%)	RR 0.75 (0.51 to 9)	77 fewer per 1000 (from 151 fewer to 1000 more)	⊕⊕○○ LOW	IMPORTANT
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None		0%	RR 0.75 (0.51 to 9)	—	⊕⊕○○ LOW	IMPORTANT

1: Allocation concealment and blinding unclear. Attrition unclear.

2: Small effect size, 95% CI includes null.

Table 9.11Should gentian violet versus ketoconazole be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Gentian violet	Ketoconazole	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure (follow-up mean 14 days)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	11/49 (22.4%)	10/45 (22.2%)	RR 1.01 (0.47 to 2.15)	2 more per 1000 (from 118 fewer to 256 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None		22.2%	RR 1.01 (0.47 to 2.15)	2 more per 1000 (from 118 fewer to 255 more)	⊕⊕○○ LOW	CRITICAL
Mycological cure (follow-up mean 14 days)
1	Randomized trial	Serious	No serious inconsistency	No serious indirectness	Serious	None	16/49 (32.7%)	13/45 (28.9%)	RR 1.13 (0.61 to 2.08)	38 more per 1000 (from 113 fewer to 312 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious	No serious inconsistency	No serious indirectness	Serious	None		28.9%	RR 1.13 (0.61 to 2.08)	38 more per 1000 (from 113 fewer to 312 more)	⊕⊕○○ LOW	CRITICAL

1: No blinding due to character of drug – dye. Small study. 95% CI includes null.

2: Small study, less than 300 events. 95% CI includes null.

Table 9.12Should gentian violet versus nystatin be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Gentian violet	Nystatin	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure (follow-up mean 14 days)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	11/49 (22.4%)	2/47 (4.3%)	RR 5.28 (1.23 to 22.55)	182 more per 1000 (from 10 more to 917 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None		4.3%	RR 5.28 (1.23 to 22.55)	184 more per 1000 (from 10 more to 927 more)	⊕⊕○○ LOW	CRITICAL
Mycological cure (follow-up mean 14 days)
1	Randomized trial	Serious	No serious inconsistency	No serious indirectness	Serious²	None	16/49 (32.7%)	3/47 (6.4%)	RR 5.12 (1.59 to 16.42)	263 more per 1000 (from 38 more to 984 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious	No serious inconsistency	No serious indirectness	Serious²	None		6.4%	RR 5.12 (1.59 to 16.42)	264 more per 1000 (from 38 more to 987 more)	⊕⊕○○ LOW	CRITICAL

1: No blinding was used.

2: Small study, less than 300 events.

Table 9.13Should ketoconazole versus nystatin be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Ketoconazole	Nystatin	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure (follow-up mean 14 days)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	10/45 (22.2%)	2/47 (4.3%)	RR 5.22 (1.21 to 22.53)	180 more per 1000 (from 9 more to 916 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None		4.3%	RR 5.22 (1.21 to 22.53)	181 more per 1000 (from 9 more to 926 more)	⊕⊕○○ LOW	CRITICAL
Mycological cure (follow-up mean 14 days)
1	Randomized trial	Serious¹^,²	No serious inconsistency	No serious indirectness	Serious	None	13/45 (28.9%)	3/47 (6.4%)	RR 4.53 (1.38 to 14.83)	225 more per 1000 (from 24 more to 883 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious¹^,²	No serious inconsistency	No serious indirectness	Serious	None		6.4%	RR 4.53 (1.38 to 14.83)	226 more per 1000 (from 24 more to 885 more)	⊕⊕○○ LOW	CRITICAL

1: No blinding was used, small study.

2: Small study, less than 300 events.

Table 9.14Should caspofungin versus amphotericin B be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Caspofungin	Amphotericin B	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure (follow-up 7–14 days)
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious¹	None	36/40 (90%)	8/12 (66.7%)	RR 1.35 (0.89 to 2.04)	233 more per 1000 (from 73 fewer to 693 more)	⊕⊕⊕○ MODERATE	CRITICAL
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious¹	None		66.7%	RR 1.35 (0.89 to 2.04)	233 more per 1000 (from 73 fewer to 694 more)	⊕⊕⊕○ MODERATE	CRITICAL

1: Small study with only 52 participants. Small estimate of effect and 95% CI includes null.

Table 9.15Should posaconazole versus fluconazole be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Posaconazole	Fluconazole	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure (follow-up mean 14 days)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	130/135 (96.3%)	139/143 (97.2%)	RR 1.32 (0.36 to 4.83)	311 more per 1000 (from 622 fewer to 1000 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None		97.2%	RR 1.32 (0.36 to 4.83)	311 more per 1000 (from 622 fewer to 1000 more)	⊕⊕○○ LOW	CRITICAL
Mycological cure (follow-up mean 14 days)
1	Randomized trial	Serious	No serious inconsistency	No serious indirectness	Serious²	None	24/91 (26.4%)	41/101 (40.6%)	RR 1.24 (1.01 to 1.52)	97 more per 1000 (from 4 more to 211 more)	⊕⊕○○ LOW	IMPORTANT
1	Randomized trial	Serious	No serious inconsistency	No serious indirectness	Serious²	None		40.6%	RR 1.24 (1.01 to 1.52)	97 more per 1000 (from 4 more to 211 more)	⊕⊕○○ LOW	IMPORTANT
Mycological eradication (follow-up mean 14 days)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious³	None	22/91 (24.2%)	36/101 (35.6%)	RR 1.18 (0.98 to 1.42)	64 more per 1000 (from 7 fewer to 150 more)	⊕⊕○○ LOW	IMPORTANT
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious³	None		35.6%	RR 1.18 (0.98 to 1.42)	64 more per 1000 (from 7 fewer to 150 more)	⊕⊕○○ LOW	IMPORTANT

1: No indication of how randomization was done. Allocation concealment not reported. Patients not blinded.

2: Small estimate of effect. 95%CI includes null. Less than 300 events.

3: No explanation was provided.

Table 9.16Should lemon juice versus gentian violet be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Lemon juice	Gentian violet	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure (follow-up mean 10 days)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	16/30 (53.3%)	9/30 (30%)	RR 1.78 (0.94 to 3.37)	234 more per 1000 (from 18 fewer to 711 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None		30%	RR 1.78 (0.94 to 3.37)	234 more per 1000 (from 18 fewer to 711 more)	⊕⊕○○ LOW	CRITICAL
Clinical failure (follow-up mean 10 days)
1	Randomized trial	Serious³	No serious inconsistency	No serious indirectness	Serious³	None	2/30 (6.7%)	8/30 (26.7%)	RR 0.25 (0.06 to 1.08)	200 fewer per 1000 (from 251 fewer to 21 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious³	No serious inconsistency	No serious indirectness	Serious³	None		26.7%	RR 0.25 (0.06 to 1.08)	200 fewer per 1000 (from 251 fewer to 21 more)	⊕⊕○○ LOW	CRITICAL

1: No blinding reported, intervention is a dye and difficult to blind.

2: Small study with less than 300 events, small estimate of effect.

3: Small study with less than 300 events, 95% CI includes null.

Table 9.17Should lemon grass versus gentian violet be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Lemon grass	Gentian violet	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure (follow-up mean 10 days)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious¹	None	15/30 (50%)	9/30 (30%)	RR 1.67 (0.87 to 3.2)	201 more per 1000 (from 39 fewer to 660 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious¹	None		30%	RR 1.67 (0.87 to 3.2)	201 more per 1000 (from 39 fewer to 660 more)	⊕⊕○○ LOW	CRITICAL
Clinical failure (follow-up mean 10 days)
1	Randomized trial	Serious	No serious inconsistency	No serious indirectness	Serious	None	2/30 (6.7%)	8/30 (26.7%)	RR 0.25 (0.06 to 1.08)	200 fewer per 1000 (from 251 fewer to 21 more)	⊕⊕○○ LOW	IMPORTANT
1	Randomized trial	Serious	No serious inconsistency	No serious indirectness	Serious	None		26.7%	RR 0.25 (0.06 to 1.08)	200 fewer per 1000 (from 251 fewer to 21 more)	⊕⊕○○ LOW	IMPORTANT

1: Small study with less than 300 events. 95% CI includes null.

Table 9.18Should lemon juice versus lemon grass be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Lemon juice	Lemon grass	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure (follow-up mean 10 days)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None	16/30 (53.3%)	15/30 (50%)	RR 1.07 (0.65 to 1.74)	35 more per 1000 (from 175 fewer to 370 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious²	None		50%	RR 1.07 (0.65 to 1.74)	35 more per 1000 (from 175 fewer to 370 more)	⊕⊕○○ LOW	CRITICAL
Clinical failure (follow-up mean 10 days)
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious	None	2/30 (6.7%)	2/30 (6.7%)	RR 1 (0.15 to 6.64)	0 fewer per 1000 (from 57 fewer to 376 more)	⊕⊕○○ LOW	CRITICAL
1	Randomized trial	Serious¹	No serious inconsistency	No serious indirectness	Serious	None		6.7%	RR 1 (0.15 to 6.64)	0 fewer per 1000 (from 57 fewer to 378 more)	⊕⊕○○ LOW	CRITICAL

1: No blinding reported.

2: Small study, few events. Small estimate of effect and 95% CI includes null.

Table 9.19Should miconazole versus clotrimazole be used for the management of oropharyngeal candidiasis associated with HIV infection in adults and children?

Quality assessment							No. of patients		Effect		Quality	Importance
No. of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Miconazole	Clotrimazole	Relative (95% CI)	Absolute	Quality	Importance
Clinical cure (follow-up mean 14 days)
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious¹	None	176/290 (60.7%)	187/287 (65.2%)	RR 0.93 (0.82 to 1.06)	46 fewer per 1000 (from 117 fewer to 39 more)	⊕⊕⊕○ MODERATE	CRITICAL
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious¹	None		65.2%	RR 0.93 (0.82 to 1.06)	46 fewer per 1000 (from 117 fewer to 39 more)	⊕⊕⊕○ MODERATE	CRITICAL
Mycological cure (follow-up mean 35 days)
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious¹	None	79/290 (27.2%)	71/287 (24.7%)	RR 1.1 (0.84 to 1.45)	25 more per 1000 (from 40 fewer to 111 more)	⊕⊕⊕○ MODERATE	IMPORTANT
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious¹	None		24.7%	RR 1.1 (0.84 to 1.45)	25 more per 1000 (from 40 fewer to111 more)	⊕⊕⊕○ MODERATE	IMPORTANT
Relapse (follow-up mean 35 days)
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious¹	None	48/172 (27.9%)	52/185 (28.1%)	RR 0.99 (0.71 to 1.38)	3 fewer per 1000 (from 82 fewer to 107 more)	⊕⊕⊕○ MODERATE	IMPORTANT
1	Randomized trial	No serious risk of bias	No serious inconsistency	No serious indirectness	Serious¹	None		28.1%	RR 0.99 (0.71 to 1.38)	3 fewer per 1000 (from 81 fewer to 107 more)	⊕⊕⊕○ MODERATE	IMPORTANT

1: Small estimate of effect with 95% CI including null.

Web Appendix 1GRADE tables

1. Kaposi sarcoma

Kaposi sarcoma in children

Mild and moderate Kaposi sarcoma

Severe or progressive Kaposi sarcoma

2. Seborrhoeic dermatitis

3. Papular pruritic eruption (PPE)

4. Eosinophilic folliculitis*

5. Tinea infections

6. Herpes zoster

7. Scabies

8. Molluscum contagiosum

9. Oropharyngeal candidiasis

10. Stevens-Johnson syndrome and toxic epidermal necrolysis

Studies not appropriate for GRADE

References for Web Appendix 1 GRADE tables1

Kaposi sarcoma