Criteria for considering studies for this review

Search methods for identification of studies

Electronic search
We will use electronic search strategies to identify relevant trials (as defined under 'type of studies'), as well as reviews/meta‐analysis (for identification of additional trials). The following databases will be searched:

The Cochrane Central Register of Controlled Trials (CENTRAL) ‐ The Cochrane Library Issue 4, 2002.
MEDLINE (January 1966‐April 2005).
EMBASE (January 1966‐April 2005).

We will also search databases for ongoing trials:

Current Controlled Trials (http://www.controlled‐trials.com)
UK National Research Register (http://www.update‐software.com/national/nrr‐frame.html)
Center Watch Clinical Trials Listing Service (http://www.CenterWatch.com/)
National Institute of Health (http://clinicalstudies.info.nih.gov/)

The following MEDLINE (OVID) search strategy will be adapted for the other electronic databases searched. The full search strategy for each database can be obtained from the Editorial Base of the Metabolic and Endocrine Disorders Group.

Diabetes mellitus
1. exp diabetes mellitus/
2. diabet$.tw.
3. IDDM.tw.
4. NIDDM.tw.
5. MODY.tw.
6. insulin$ secret$ dysfunc$.tw.
7. impaired glucose toleran$.tw.
8. exp glucose intolerance/
9. glucose intoleran$.tw.
10. exp insulin resistance/
11. insulin$ resist$.tw.
12. (non insulin$ depend$ or noninsulin$ depend$ or non insulin?depend$ or noninsulin?depend$).tw.
13. (insulin? depend$ or insulin?depend$).tw
14. metabolic$ syndrom$.tw.
15. (pluri metabolic$ syndrom$ or plurimetabolic$ syndrom$).tw.
16. or/1‐15
17. exp diabetes insipidus/
18. diabet$ insipidus.tw.
19. 17 or 18
20. 16 not 19

Erectile dysfunction
1. exp Impotence/
2. erecti$ dysfunct$.mp.
3. sex$ dysfunct$.mp.
4. sex$ disorder$.mp.
5. exp PENILE ERECTION/
6. erect$.tw.
7. 1 or 2 or 3 or 4 or 5 or 6

Phosphodiesterase type 5 inhibiors
1. exp 3',5'‐cyclic‐gmp phosphodiesterase/
2. phospho?di?esterase$.mp
3. PDE?.mp
4. sildenafil$.tw
5. viagra$.tw
6. vardenafil$.tw
7. levitra$.tw
8. tadalafil$.tw
9. cialis$.tw
10. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9

Controlled or randomised clinical trials

Phase I
1. randomised controlled trial.pt.
2. controlled clinical trial.pt.
3. Randomised Controlled Trials/
4. Random Allocation/
5. Double‐Blind Method/
6. Single Blind Method/
7. 1 or 2 or 3 or 4 or 5 or 6
8. Animal/ not Human/
9. 7 not 8

Phase II
10. clinical trial.pt.
11. exp Clinical Trials/
12. (clinic$ adj25 trial$).tw.
13. ((singl$ or doubl$ or trebl$ or tripl$) adj (mask$ or blind$)).tw.
14. Placebos/
15. placebo$.tw.
16. random$.tw.
17. Research Design/
18. (latin adj square).tw.
19. 10 or 13 or 12 or 13 or 14 or 15 or 16 or 17 or 18
20. 19 not 8
21. 20 not 9

Phase III
22. Comparative Study/
23. exp Evaluation Studies/
24. Follow‐Up Studies/
25. Prospective Studies/
26. (control$ or prospectiv$ or volunteer$).tw.
27. Cross‐Over Studies/
28. 22 or 23 or 24 or 25 or 26 or 27
29. 28 not 8
30. 29 not (9 or 21)

All phases
33. 9 or 21 or 30

Meta analysis or systematic reviews
1. exp meta‐analysis/
2. exp Review Literature/
3. meta‐analysis.pt.
4. review.pt.
5. 1 or 2 or 3 or 4
6. letter.pt.
7. comment.pt.
8. editorial.pt.
9. historical‐article.pt.
10. 6 or 7 or 8 or 9
11. 5 not 10
12. ((systematic$ or quantitativ$ or methodologic$) adj (review$ or overview$)).tw.
13. meta?anal$.tw.
14. (integrativ$ research review$ or research integration$).tw.
15. quantitativ$ synthes$.tw.
16. (pooling$ or pooled analys$ or mantel$ haenszel$).tw.
17. (peto$ or der?simonian$ or fixed effect$ or random effect$).tw.
18. 12 or 13 or 14 or 15 or 16 or 17
19. 11 or 18
20. limit 19 to human [Limit not valid in: Pre‐MEDLINE; records were retained]

Note: (ADJn) retrieves two or more query terms within n words of each other, and in any order; a question mark (?) can be used within or at the end of a query word to substitute for one or no characters; a truncation mark ($) stands for unlimited truncation to retrieve all possible suffix variations of a root word; (exp)‐ all MeSH terms and a specific category will be included in the search; MeSH: Medical Subject Heading; (mp) looks for exact word or phrase in the title, abstract and subject‐heading fields; (tw) looks for exact word or phrase in the title and abstract.

Handsearching
We will try to identify additional studies by searching the reference lists of relevant trials and reviews identified.

Other search strategies
We will contact the manufacturers of sildenafil (Viagra), vardenafil (Levitra) and tadalafil (Cialis) (Pfizer, Bayer, Lilly, respectively). In addition, we will try to identify and contact researchers conducting ongoing trials.

Search for identification of studies will not be restricted by language.

Data collection and analysis

Trial Selection
Two independent reviewers (MV, AN) will scan the titles, abstract sections and keywords of every record retrieved. Full articles will be retrieved for further assessment if the information given suggests that the study fulfills the insertion criteria and does not meet the exclusion criteria. If there is any doubt regarding these criteria from the information given in the title and abstract, the full article will be retrieved for clarification. Interrater agreement for study selection will be measured using the kappa statistic (Cohen 1960). Where differences in opinion exist, they will by resolved through open discussion. In the case of further dispute, the article will be added to those 'awaiting assessment' and the authors will be contacted for clarification. If no clarification is provided, the review group editorial base will be consulted. The level of interrater agreement for study selection will be reported.

Quality assessment of trials
The quality of reporting each trial will be assessed based largely on the quality criteria specified by Schulz and by Jadad (Jadad 1996; Schulz 1995). In particular, the following factors will be studied:
1. Minimisation of selection bias ‐ was the randomisation procedure adequate? was the allocation concealment adequate?
2. Minimisation of performance bias ‐ were the patients and people administering the treatment blind to the intervention?
3. Minimisation of attrition bias ‐ were withdrawals and dropouts completely described? was analysis by intention‐to‐treat?
4. Minimisation of detection bias ‐ were outcome assessors blind to the intervention?

Based on these criteria, studies will be broadly subdivided into the following three categories:
(A) all quality criteria met: low risk of bias;
(B) one or more of the quality criteria only partly met: moderate risk of bias; and
(C) one or more criteria not met: high risk of bias.

This classification will be used as the basis of a sensitivity analysis. Additionally, we will explore the influence of individual quality criteria in a sensitivity analysis.

Two reviewers (MV, AN) will independently assess each trial. Interrater agreement will be calculated using the kappa statistic, and reported. In cases of disagreement, judgment will be made based upon consensus.

Data Extraction
A data extraction form is used and submitted with this protocol. The following data will be extracted.
1. General information: author, title, publication (type, unpublished), language of publication, year of publication, country, complete reference or source, contact details, duplicate publication, multiple publication, rural or city, single centre versus multicentre, setting, stated aim of the study, sponsor, ethic committee approval and description of conflict of interests.
2. Trial design: prospective study, control group, parallel study, placebo controlled, active medication controlled, cross‐over study, run‐in period, wash‐out period (for cross‐over trials), carryover effect described (for cross‐over trials), period effect described, sampling method, power calculation, selection bias (randomisation, unit of randomisation and allocation concealment adequacy), performance bias (blinding of patients and caregivers, method of blinding, check of blinding, check of blinding method), attrition bias (intention‐to‐treat analysis, withdrawals description, drop‐outs description, losses to follow‐up description, change of groups (cross‐overs), number of drop‐out and withdrawals and losses to follow‐up, reasons for drop‐outs or withdrawals or losses to follow‐up description), detection bias (blinding of outcome assessors), overall quality assessment, definition of inclusion criteria, definition of exclusion criteria, specification of exclusion criteria, predefined subgroups, posthoc defined subgroups and specification of subgroups.
3. Participants: diabetes mellitus diagnostic criteria description, diabetes mellitus diagnostic criteria validity, erectile dysfunction diagnostic criteria description, erectile dysfunction diagnostic criteria validity, exclusion criteria definition, baseline characteristics i.e. age, gender, race, body mass index, glycated haemoglobin, fasting plasma glucose, duration of diabetes mellitus, type of diabetes mellitus, diabetes mellitus related complications, diabetes mellitus related treatment, comorbidities, other medications, identical treatment of groups (apart from intervention), prior treatment for erectile dysfunction and baseline erectile dysfunction status.
4. Intervention: nature of therapy, regimen (dose, schedule and units), duration of therapy, length of follow‐up, compliance
5. Outcomes: total deaths, adverse events, efficacy as measured with regards to erectile dysfunction status.
6. Effect modifiers: glycated haemoglobin, fasting plasma glucose, compliance, change of concomitant medication.

Both data extraction and data entry will be done in duplicate by two reviewers (MV, AN). Differences in data extraction will be resolved by consensus after referring back to the original paper.

If necessary, we will contact the authors of studies about missing information in their trials.

Data analysis
Data will be summarised statistically if they are available, sufficiently similar and of sufficient quality.

Data type
The outcome data is expected to be of ordinal nature (including measurement scales), where the outcome is one of several ordered categories or generated by scoring and summing categorical responses (such as in the case of self reported quality of life assessment scale). We also anticipate some data to be presented as counts‐ e.g. number of satisfying penetrations per time period.

Measures of effect size
Effect measures for ordinal data is intended to be assessed as continuous data, unless the ordinal scale will be short. In the former case, standard mean difference will be measured, to enable standardization of the results of different trials into a uniform scale before combining them in a meta‐analysis. This method will be used for each of the parameters assessed, after assuring that the different scales used in the trials point towards a single direction. If this is not the case, scale direction will be reversed by multiplying the mean values by (‐1). It is acknowledged that the use of standard mean difference rather than weighted mean difference may produce difficulties in interpreting results and effort will be made to re‐express these results back to clinical scales. Whenever dichotomizing results will seem statistically feasible and clinically logical, we will use effect measurements for dichotomous data, i.e. risk ratio and odds ratio.

Effect measures for count and rates data will be assessed as continuous data given the anticipation for common event occurrence. Weighted mean difference will be used to assess the difference in the mean number of events per group of patients per unit of time.

Short‐, intermediate‐ and long‐term assessment of effect in the case of repeated observations per participant do not seem clinically significant in our review. Therefore, in the case of multiple observations we will select the longest follow‐up for each trial and analyse results to that point. This method may induce lack of consistency across trials and give rise to heterogeneity, which will be addressed in the interpretations of results.

intention‐to‐treat analysis aims to include all participants randomised into a trial irrespective of what happened subsequently. This means that ideally, trial participants should be analysed in the group to which they were randomised regardless of which treatment they actually received or other protocol irregularities and all participants should be included regardless of whether their outcomes were actually collected. In this review we adopt an available‐case‐analysis, in which only the former criterion is met, and analysis is performed for every participant for which data is available, thus filling‐in for missing data will not be conducted. Special attention will be given for three types of exclusions:
(a) participants excluded for predefined exclusion criteria using information collected before randomisation‐ will be considered as legitimate
(b) participants excluded immediately after randomisation‐ illegitimate
(c) very high dropout rates or inconsistency across study groups, which may indicate low quality of trial conduction.

Heterogeneity
Heterogeneity is a term used to describe variability among trials encountered in a meta‐analysis. In our review, we anticipate some clinical diversity, e.g. diabetes status, erectile dysfunction status or treatment, as well as methodological diversity, both giving rise to statistical heterogeneity. On the one hand, the scope of our review is relatively confined, therefore heterogeneity is not inherently large. On the other hand, disease status may be variable among participants and treatment in the study group relates to three different drugs, while comparison groups are even more diverse. Methodologically there is no promise for comparable quality. In this review, it seems appropriate to aim at broader perspective, thus, primarily evaluate the average effect of phosphodiesterase type 5 inhibitors by combining results from trials that evaluate the different drugs under this pharmacological class.

Heterogeneity will be identified using the formal chi‐square test, which is intended to assess whether observed differences in results are compatible with chance alone. Chi‐squared test have low power when trials have small sample size or are few in number. Therefore a P value of 0.10 will be considered statistically significant. We will also try and quantify the amount of heterogeneity by describing the percentage of the variability in effect estimates that is due to heterogeneity rather than chance (Higgins 2002; Higgins 2003).

If heterogeneity will prove to be substantial, it may be unwise to perform a meta‐analysis.

A negative chi‐square will not necessarily mean that no heterogeneity exists or should not be further explored. In this view, a subgroup analysis will be performed as discussed below.

Meta‐analytic model
Heterogeneity will be addressed using two meta‐analytical methods. In the case of limited number of studies retrieved, a fixed‐effect meta‐analysis will be carried out, ignoring heterogeneity, thus calculating the typical group treatment effect. Fixed‐effect analysis will be performed using the Mantel‐Haenszel methods, which have been shown to have better statistical properties when there are few events. If sufficient amount of studies are found, a random‐effect meta‐analysis will be performed. This will be conducted with DerSimonian and Laird method.

When addressing dichotomous data types, we will express the effect of treatment with relative effect measures, i.e. risk ratio and odds ratio, which are more consistent than absolute measures. These will be calculated for an event (rather then for a non‐event) and re‐expressed as absolute measures which are more interpretable by clinicians. When addressing continuous data type, we will express the effect of treatment using standardised mean difference for ordinal data type and weighted mean difference for counts data type.

Subgroup analysis
To further explore heterogeneity and investigate the effect modification of participants and treatment types, we will perform a subgroup analysis, according to the following clinically logical predefined groups. If more than ten articles are available for each of the following characteristics, a meta‐regression model will be used. Otherwise, standard subgroup analysis will be followed. It is noted that if additional meaningful characteristics will be found, a posthoc analysis will be conducted.

1. Participants:
1.a. gender‐ male versus female
1.b. diabetes status‐ mild to moderate versus severe (according to clinical status and treatment).
1.c. erectile dysfunction status‐ mild to moderate versus severe (according to baseline status).
2. Intervention:
2.a. Phosphodiesterase type 5 inhibitors: Viagra versus Cialis versus Levitra

A dose‐response analysis will not be performed, nor any indirect comparisons between groups not directly evaluated head to head in a clinical trial.

Sensitivity analysis
We will perform sensitivity analyses in order to explore the influence of the following factors on effect size by:
1. repeating the analysis excluding unpublished studies (if there were any)
2. repeating the analysis taking account of study quality (low‐, moderate‐ or high‐risk of bias)
3. repeating the analysis taking account of specific quality criteria (i.e. selection, performance, attrition and detection biases)
4. repeating the analysis excluding any very long or large studies to establish how much they dominate the results
5. repeating the analysis excluding studies using the following filters: diagnostic criteria for diabetes mellitus and erectile dysfunction, language of publication, source of funding (industry versus other), country and scales used for measuring effect (validated versus other).

The robustness of the results will also be tested by repeating the analysis using different measures of effects size (risk difference, odds ratio etc.) and different statistic models (fixed and random effects models).

Assessment of publication bias
Publication bias will be analysed with the funnel plot method. Funnel plot asymmetry would be assessed statistically. It is, however, important to realise that publication bias is only one of a number of possible causes of funnel‐plot asymmetry.