Figure 1. Literature search flow diagram. CENTRAL indicates Cochrane Central Register of Controlled Trials.
Figure 2. Galbraith plot. There are 2 statistical outliers, A and B, which represent the trials by Ferrara et al25 and Barbosa-Cesnik et al,26 respectively.
Figure 3. Influential plot: pooled summary effect estimates with each study omitted at one time.
Figure 4. Forest plot: summary effect of cranberry in prevention of urinary tract infection, expressed as risk ratio (RR). W(fixed) indicates weights in fixed-effect Mantel-Haenszel model.
Figure 5. Funnel plot to evaluate publication bias.
Wang C-H, Fang C-C, Chen N-C, et al. Cranberry-containing products for prevention of urinary tract infections in susceptible populations: a systematic review and meta-analysis of randomized controlled trials. Arch Intern Med. doi:10.1001/archinternmed.2012.3004
eAppendix. Search strategy for MEDLINE and detailed list of variables for data extraction and quality assessment
Wang C, Fang C, Chen N, Liu SS, Yu P, Wu T, Chen W, Lee C, Chen S. Cranberry-Containing Products for Prevention of Urinary Tract Infections in Susceptible PopulationsA Systematic Review and Meta-analysis of Randomized Controlled Trials. Arch Intern Med. 2012;172(13):988–996. doi:10.1001/archinternmed.2012.3004
Author Affiliations: Department of Emergency Medicine, National Taiwan University Hospital, and National Taiwan University College of Medicine, Taipei (Drs Wang, Fang, N-C Chen, Liu, Yu, Wu, W-T Chen, and S-C Chen); and Department of Emergency Medicine, National Taiwan University Hospital Yunlin Branch, Yunlin, and Department of Epidemiology, Harvard School of Public Health, Boston, Massachusetts (Dr Lee).
Background Urinary tract infection (UTI) is one of the most commonly acquired bacterial infections. Cranberry-containing products have long been used as a folk remedy to prevent UTIs. The aims of this study were to evaluate cranberry-containing products for the prevention of UTI and to examine the factors influencing their effectiveness.
Methods MEDLINE, EMBASE, and the Cochrane Central Register of Controlled Trials were systemically searched from inception to November 2011 for randomized controlled trials that compared prevention of UTIs in users of cranberry-containing products vs placebo or nonplacebo controls. There were no restrictions for language, population, or publication year.
Results Thirteen trials, including 1616 subjects, were identified for qualitative synthesis from 414 potentially relevant references; 10 of these trials, including a total of 1494 subjects, were further analyzed in quantitative synthesis. The random-effects pooled risk ratio (RR) for cranberry users vs nonusers was 0.62 (95% CI, 0.49-0.80), with a moderate degree of heterogeneity (I2 = 43%) after the exclusion of 1 outlier study. On subgroup analysis, cranberry-containing products seemed to be more effective in several subgroups, including women with recurrent UTIs (RR, 0.53; 95% CI, 0.33-0.83) (I2 = 0%), female populations (RR, 0.49; 95% CI, 0.34-0.73) (I2 = 34%), children (RR, 0.33; 95% CI, 0.16-0.69) (I2 = 0%), cranberry juice drinkers (RR, 0.47; 95% CI, 0.30-0.72) (I2 = 2%), and subjects using cranberry-containing products more than twice daily (RR, 0.58; 95% CI, 0.40-0.84) (I2 = 18%).
Conclusions Our findings indicate that cranberry-containing products are associated with protective effect against UTIs. However, this result should be interpreted in the context of substantial heterogeneity across trials.
Urinary tract infection (UTI) is defined by the presence of clinical signs and symptoms arising from the genitourinary tract in the presence of a certain threshold level of bacteria in the urine. It is one of the most commonly acquired bacterial infections in ambulatory and hospitalized populations.1 It is estimated that UTIs cause approximately 7 000 000 office visits each year in the United States, along with an additional 1 000 000 visits to emergency departments, and result in approximately 100 000 hospitalizations.1 The estimated annual cost of community-acquired UTIs is $1.6 billion in the United States.2
Adult women are particularly susceptible to UTIs. Up to 40% to 50% of women will experience at least 1 episode of UTI during their lifetimes.2 Approximately 20% to 30% of women who have had an infection will experience a recurrence with concurrent short-term morbidity.2 Other groups with an increased risk of UTI include pregnant women, elderly patients, and patients with neuropathic bladder.3
Cranberry (genus Vaccinium, including the species V oxycoccus, V macrocarpon, V microcarpum, and V erythrocarpum) has been used as a folk remedy to prevent UTI for many years.4 In the 1920s, the preventive effect of cranberry was thought to be a result of acidification of the urinary tract, but this was refuted in 1959.3 In 1984, cranberry was found to interfere with the attachment of bacteria to uroepithelial cells, thereby potentially preventing infection.3 In 1989, A-type proanthocyanidins (PACs) were identified as compounds with the potential to inhibit the adherence of P-fimbriated Escherichia coli to the urogenital mucosa,5 and there are hundreds of other compounds found in cranberries that have yet to be explored for their potential antiadherence activity.
A Cochrane meta-analysis by Jepson and Craig6 showed some evidence that cranberry juice decreases the number of symptomatic UTIs, particularly among women with recurrent UTIs. However, because of a limited number of included trials, it did not explore how other factors, such as dose, frequency, or form of cranberry-containing products, influence the effectiveness in prevention of UTI. Furthermore, since its publication, several new studies have been published. The goals of the present review were to reevaluate cranberry-containing products for the prevention of UTI and to examine the factors influencing their effectiveness.
We performed this meta-analysis in accordance to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guidelines for systematic reviews.7 Two authors (C.-H.W. and C.-C.F.) independently searched MEDLINE, EMBASE, and the Cochrane Central Register of Controlled Trials (CENTRAL) from inception to November 2011 according to a prespecified protocol. Search terms included cranberry, Vaccinium macrocarpon, Vaccinium oxycoccus, Vaccinium microcarpum, Vaccinium erythrocarpum, Vaccinium, urinary tract infection, pyelonephritis, cystitis, bacteriuria, and pyuria. No language, population, or publication year restrictions were enforced. The search strategy for MEDLINE is presented as an example in the eAppendix. We did not search abstracts from conferences, proceedings, or clinical trial registries. Instead, we manually checked bibliographies of relevant studies, reviews, and meta-analyses to identify references that may have been missed in our primary search. Authors of included trials were contacted for missing information if necessary.
Two authors (C.-H.W. and C.-C.F.) independently scanned the titles and abstracts of all retrieved manuscripts to identify those pertinent to this review. The following prespecfied inclusion criteria were used: (1) randomized controlled trials (RCTs); (2) comparison of cranberry-containing products vs placebo or nonplacebo control for prevention of UTI; and (3) outcomes reported as incidence of UTIs. (One trial8 was included according to this criterion but not pooled in the quantitative analysis because it reported incidence in a manner that we could not pool its with others.)
After retrieving full reports of potentially relevant trials, the same reviewers independently assessed the eligibility of the studies on the basis of the inclusion criteria and settled differences of opinion by consensus or by consultation with a third investigator (C.-C.L.).
Four authors (N.-C.C., S.S.-H.L., P.-H.Y., and T.-Y.W.) independently extracted data by a prespecified protocol. Each trial was reviewed independently by any 2 of the 4 authors to ensure the correctness of the extracted data. The information included the following: (1) type of study and study design; (2) characteristics of the study population; (3) types of intervention and controls; (4) definitions of UTI; (5) types of outcomes measured; and (6) number and reasons of participants lost to follow-up. The detailed list of variables is presented in the eAppendix.
The primary outcome was prespecified as the incidence of UTI, which was expressed as either incidence or cumulative incidence rate, depending on the original data. If the outcome data could not be expressed as such, the original data were presented.
The Cochrane risk of bias tool was adopted to assess the risk of bias for each trial, which was scored as “high risk,” “low risk,” or “unclear” based on methods of random sequence generation, allocation concealment, blinding process, incomplete outcome data, and selective reporting.9 Discrepancies in assessment were resolved through discussions between the authors (C.-H.W. and C.-C.F.) or consultation with a third investigator (C.-C.L.).
Data synthesis and analysis were performed using the metafor package in the R 2.11.1 statistical software (R Foundation for Statistical Computing). P ≤ .05 (2-sided) was considered statistically significant. Dichotomous outcomes from individual studies were collected to compute individual-study risk ratios (RRs) with 95% confidence intervals. To deal with studies with multiple intervention arms, all relevant treatment arms were grouped together by adding together the sample sizes and numbers of subjects with events.10 Heterogeneity was tested using both the I2 statistic, with I2 of 50% or greater indicating a substantial level of heterogeneity, and the statistical test of heterogeneity, with P ≤ .05 indicating heterogeneity.11,12 Random-effects summary estimates (DerSimonian-Laird method)13 were reported if the heterogeneity was significant in either test of heterogeneity; otherwise, fixed-effect summary estimates (Mantel-Haenszel method) were reported.
We used a Galbraith plot to identify potential sources of heterogeneity and an influential plot to assess the impact on pooled summary estimates when significant heterogeneity was detected.14 We conducted sensitivity analyses based on risks of bias in the randomization process, study characteristics, and definitions of UTI. We performed subgroup analyses by prespecified covariates, including study population, sex, age, and form, amount, and frequency of cranberry-containing products used, and tested these covariates for significance by meta-regression. We drew a funnel plot to evaluate publication bias.
In this systematic review, we identified 13 RCTs that were eligible for qualitative synthesis (Figure 1).8,15- 26 There were 9 parallel-group and 4 crossover trials (Table 1). None of these crossover trials had washout periods. Eight trials were conducted according to the intention-to-treat principle, and 5 trials used per-protocol analysis. Most of the trials did not report their randomization processes adequately and suffered from a high proportion of subjects lost to follow-up (0%-48%).
There were 1616 subjects included in the qualitative analysis (Table 2). Of the 13 trials, 10 were performed in North America (United States and Canada), and the other 3 were conducted in Europe (United Kingdom, Finland, and Italy). All but 1 of the trials22 followed subjects living in the community. According to the inclusion and exclusion criteria of each trial, we further categorized each study population into the following subgroups for subgroup analysis: women with recurrent UTIs, elderly patients, patients with neuropathic bladder, pregnant women, and children.
Administration of cranberry-containing products differed significantly in form, daily dosage, PAC content, and dosing frequency (Table 2). Nine trials used cranberry juice, and 4 used cranberry capsules or tablets. Six trials used cranberry-containing products provided by the manufacturer Ocean Spray. Daily cranberry dose ranged from 0.4 to 194.4 g. The daily use of cranberry-containing products was reported in 3 trials without specifying the actual cranberry amount.8,17,20 Cranberry-containing products were administered for 6 months in most trials.8,15,18,19,21,23- 26
Of the 3 trials involving pediatric groups,15,17,25 only Foda et al15 adjusted the daily cranberry dosage by body weight. Ten trials used a formulated placebo, 2 trials18,25 did not use a placebo, and 1 trial15 used water as the placebo.
The definitions of UTI differed significantly in thresholds for bacteriuria and pyuria (Table 3). The presence of UTI symptoms was not required in 2 trials.8,19 Because there was no reliable bioassay to examine the compliance of subjects, most studies used indirect methods for this purpose. These included periodic interviews, self-reported questionnaires, and pill counting of remaining study medications. In most trials, the occurrence of UTI was expressed as incidence or cumulative incidence rate. In 1 trial,8 however, the final result included baseline urine culture data, which precluded transformation of the outcome data into incidence or cumulative incidence rate. Cumulative incidence of UTI was reported in 10 trials,17- 26 which were further analyzed in quantitative synthesis.
There were 1494 subjects across the 10 trials17- 26 included in the quantitative data synthesis, with 794 in the cranberry group and 700 in the control group. There was significant heterogeneity among trials (RR, 0.68; 95% CI, 0.47-1.00) (I2 = 59%). The Galbraith plot indicated that the trials by Ferrara et al25 and Barbosa-Cesnik et al26 were potential sources of heterogeneity (Figure 2). Influential plot further demonstrated that the trial by Barbosa-Cesnik et al26 had the most significant impact on the pooled summary estimate (Figure 3), justifying its exclusion from the main analysis. After exclusion of this trial, heterogeneity decreased, and cranberry-containing products seemed to be effective in prevention of UTIs (RR, 0.62; 95% CI, 0.49-0.80) (I2 = 43%) (Figure 4).
Sensitivity analysis showed that the pooled summary estimate was stable to risks of bias in random sequence generation, study characteristics, and definitions of UTI. However, the protective effect was much stronger in 2 studies18,25 without placebo in the control group (pooled RR, 0.36; 95% CI, 0.21-0.62) (Table 4).
Subgroup analysis demonstrated that cranberry-containing products seemed to be more effective in women with recurrent UTIs (RR, 0.53; 95% CI, 0.33-0.83) (I2 = 0%), female populations (RR, 0.49; 95% CI, 0.34-0.73) (I2 = 34%), children (RR, 0.33; 95% CI, 0.16-0.69) (I2 = 0%), cranberry juice users (RR, 0.47; 95% CI, 0.30-0.72) (I2 = 2%), and people using cranberry-containing products more than twice daily (RR, 0.58; 95% CI, 0.40-0.84) (I2 = 18%), although the P values were not significant in meta-regression (Table 5). The funnel plot did not indicate significant publication bias (Figure 5).
In a 2008 Cochrane review, Jepson et al6 reported a favorable effect of cranberry juice in the prevention of symptomatic UTIs (RR, 0.66; 95% CI, 0.47-0.92 [4 trials]), especially in women with recurrent UTIs (RR, 0.61; 95% CI, 0.40-0.91 [2 trials]). In light of newly available studies, the present review has reevaluated the effectiveness of cranberry-containing products in the prevention of UTI and factors influencing their effectiveness. Our result was high in heterogeneity but included a greater number of available trials in analysis and showed a similar but nonsignificant pooled result (RR, 0.68; 95% CI, 0.47-1.00) (I2 = 59%). After exclusion of a potential outlier by Barbosa-Cesnik et al,26 our results were more similar to those of Jepson and Craig6 (RR, 0.62; 95% CI, 0.49-0.80) (I2 = 43%).
The trial by Barbosa-Cesnik et al26 showed that cranberry juice gave no protection against the risk of recurring UTI among college-aged women (RR, 1.43; 95% CI, 0.87-2.33), which made it an outlier among the trials. There are 2 possible reasons for this deviation. First, Barbosa-Cesnik et al26 used the lowest bacteriuria threshold (1000 colony-forming units [CFU]/mL) among all included trials to define UTI. Kontiokari et al,18 who used 10 000 CFU/mL as the threshold, showed a protective effect for recurrent UTI (RR, 0.44; 95% CI, 0.21-0.93) in a similar population. Second, the incidence rate of UTIs observed in the control group in the trial by Barbosa-Cesnik et al26 was almost half that in the trial by Kontiokari et al.18 Barbosa-Cesnik et al26 suggested that this was because of the inadvertent addition of ascorbic acid, which may also prevent UTI, to the placebo, along with better hydration in their control group. Whether the population studied in the trial by Barbosa-Cesnik et al26 was significantly different in other aspects from the populations in other trials could not be determined in the present meta-analysis.
Our sensitivity analyses showed that the protective effect of cranberry-containing products was stronger in nonplacebo-controlled trials, which suggest that expectations of efficacy may have played a role. The 2 trials without placebos were relatively small, and the pooled effect estimate did not change significantly after removal of these 2 trials (RR, 0.73; 95% CI, 0.55-0.97) (I2 = 25%).
By including a greater number of available studies than were available for the review by Jepson and Craig,6 other factors potentially influencing the effect of cranberry were available for subgroup analysis. The results revealed that subgroups of younger age, female sex, individuals with recurrent UTI history, and using cranberry juice rather than capsules or tablets might benefit more from cranberry-containing products than others, although the difference was not statistically significant.
Cranberry-containing products were more effective in women with recurrent UTIs than others (RR, 0.61; 95% CI, 0.40-0.91) (I2 = 0%), which was similar to the result of the review by Jepson and Craig.6 Furthermore, when results for girls25 and pregnant women24 were pooled together with those for women with recurrent UTIs,18,20 the summary estimate for this group composed of 100% female participants also demonstrated a preventive effect of cranberry-containing products (RR, 0.49; 95% CI, 0.34-0.73) (I2 = 34%).
Cranberry juice was noted to be more effective than cranberry capsules or tablets in subgroup analysis. Since the volume of cranberry juice may influence the incidence of UTI,27 this result might be because subjects drinking cranberry juice were better hydrated than those taking cranberry capsules or tablets. Alternatively, because the mechanism of the protective effect of cranberries against UTIs has not yet been elucidated, the better preventive effect of cranberry in juice form might come from the additive or synergistic effect of other unknown substances in the juice, which are devoid in cranberry capsules and tablets. However, ingesting a large volume of cranberry juice with high sugar content might raise concerns about sugar control in diabetic patients, and it may cause severe gastrointestinal upset or other adverse effects, as observed by Wing et al,24 who had to change their protocol to allow less frequent dosing to maintain compliance and avoid withdrawal of participants. Nonetheless, although these adverse effects are a concern, until the protective mechanism is clearly understood, use of cranberry in juice form might be favorable to cranberry in capsules or tablets.
A better preventive effect from cranberry-containing products was noted with dosing frequency more than twice daily. Because in vitro data have suggested that the antiadhesion activity of cranberry juice on fimbriated E coli lasts for approximately 8 hours after ingestion,3 dosing more frequently than twice daily may be a reasonable choice.
Among all included studies, only Wing et al24 attempted a dose-response study to determine the optimal dose.24 The results showed a trend toward fewer UTIs in women given higher doses of cranberry juice vs those who received placebo. Our study did not have the power to examine differences in incidence of UTI by cranberry dose. Most researchers for the trials included in the present review did not state their rationale in selecting a cranberry dose, except for McGuinness et al,19 who mentioned that the they chose the highest available cranberry dose for their study. More dose-response studies are needed to find the optimal in vivo dosage of cranberry; an ongoing study to determine dose is currently under way (clinicaltrials.gov Identifier: NCT00100061).
In trials using daily cranberry amounts greater than 10 g, a huge range of PAC contents, from 0.838 mg to 224 mg per day, was reported. In contrast, those using daily cranberry amounts of less than 10 g did not include information about PAC content. Only recently have PACs been recognized as potential antiadhesion constituents of cranberry.5 Quantification and standardization of PAC content was available in only 3 of the 13 trials included, and it is difficult to effectively compare outcomes among the trials without this information. The results of 1 in vitro study have suggested that daily intake of 72 mg of PACs may protect against bacterial adhesion in the urinary tract after cranberry consumption.28 However, in 2 trials included in the present review, using daily PAC amounts above 72 mg did not show a protective effect.24,26
Theories about the role of cranberry in the prevention of UTIs have gradually evolved and in part have only recently been formulated. The PAC content should be specified in future trials to allow differentiation of the effects of PACs on clinical response. Also, consumers of cranberry-containing products should have access to this information when they buy products touted for prevention of UTI.
Regarding study limitations, we did not search abstracts from conferences, proceedings, or clinical trial registries, which may cause incomplete literature search. We tried to compensate for this by manually checking bibliographies of relevant studies, reviews, or meta-analyses. Also, we were unable to reach several authors for missing data necessary for synthesis.8,15,16 We checked relevant meta-analyses and systemic reviews and confirmed that no such data were available.
In conclusion, the results of the present meta-analysis support that consumption of cranberry-containing products may protect against UTIs in certain populations. However, because of the substantial heterogeneity across trials, this conclusion should be interpreted with great caution. Cranberry-containing products tend to be more effective in women with recurrent UTIs, female populations, children, cranberry juice drinkers, and people using cranberry-containing products more than twice daily.
Correspondence: Chien-Chang Lee, MD, MSc, Department of Emergency Medicine, National Taiwan University Hospital Yunlin Branch #579, Sec 2, Yunlin Rd, Douliou 640, Yunlin, Taiwan (firstname.lastname@example.org).
Accepted for Publication: May 5, 2012.
Author Contributions: Dr Lee had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Fang, Lee, and S.-C. Chen. Acquisition of data: Wang, N.-C. Chen, Liu, Yu, Wu, and W.-T. Chen. Analysis and interpretation of data: Lee. Drafting of the manuscript: Wang, N.-C. Chen, Liu, Yu, Wu, W.-T. Chen, and S.-C. Chen. Critical revision of the manuscript for important intellectual content: Fang and Lee. Statistical analysis: Lee. Administrative, technical, and material support: Wang, Fang, N.-C. Chen, Liu, Yu, Wu, W.-T. Chen, and Lee. Study supervision: S.-C. Chen.
Financial Disclosure: None reported.