Received: 22 September 2016 Revised: 1 November 2016 Accepted: 9 November 2016

DOI 10.1111/dom.12825

ORIGINAL ARTICLE

Urinary tract and genital infections in patients with type 2

diabetes treated with sodium-glucose co-transporter 2
inhibitors: A meta-analysis of randomized controlled trials
Dandan Li MSc1 | Tiansheng Wang PharmD2,3 | Su Shen BSc1 | Zhenwei Fang MSc4 |
Yue Dong MSc5 | Huilin Tang MSc6,7

1
Department of Pharmacy, Beijing Friendship
Hospital, Capital Medical University, Beijing, Aims: To evaluate the effects of different sodium-glucose co-transporter 2 (SGLT2) inhibitors
China on the risk of urinary tract infections (UTIs) and genital infections in patients with type 2 diabe-
2
Department of Pharmacy Administration and tes mellitus (T2DM).
Clinical Pharmacy, Peking University Health
Materials and methods: We systematically searched PubMed, Embase, CENTRAL, and Clinical-
Science Center, Beijing, China
3
Trials.gov from inception to October 9, 2016 to identify randomized controlled trials (RCTs)
Department of Epidemiology, Gillings School
of Global Public Health, University of North reporting the occurrence of UTIs and genital infections in patients with T2DM treated with
Carolina at Chapel Hill, Chapel Hill, North SGLT2 inhibitors. Pairwise and network meta-analyses were performed to calculate odds ratios
Carolina (ORs) and 95% confidence intervals (CIs). Meta-regression was performed to assess explanatory
4
Department of Pharmacy, Beijing Anzhen factors that might influence effect size.
Hospital, Capital Medical University, Beijing,
China
Results: A total of 52 RCTs involving 36 689 patients were eligible for our meta-analysis. Cana-
5 gliflozin, dapagliflozin and empagliflozin were associated with a higher risk of genital infections
Department of Pharmacy, Huashan Hospital,
Fudan University, Shanghai, China than placebo, with ORs ranging from 3.21 (95% CI 2.08-4.93) for dapagliflozin 2.5 mg to 5.23
6
Department of Pharmacy, Peking University (95% CI 3.86-7.09) for canagliflozin 300 mg. Only dapagliflozin 10 mg led to significantly more
Third Hospital, Beijing, China UTIs than placebo. The increased risk of UTIs and genital infections seemed to have a dose–
7
Department of Epidemiology, Richard response relationship for dapagliflozin only. No factors that had a significant modification effect
M. Fairbanks School of Public Health, Indiana
on these infectious events were detected in meta-regression analysis.
University, Indianapolis, Indiana
Conclusions: The present study found that canagliflozin, dapagliflozin and empagliflozin were
Correspondence
Huilin Tang MSc, Department of Epidemiology, associated with a significantly higher risk of genital infections compared with placebo and other
Richard M. Fairbanks School of Public Health, active treatments. Only dapagliflozin had a dose–response relationship with UTIs and genital
Indiana University, 1050 Wishard Blvd,
infections.
Indianapolis, IN 46202.
Email: huiltang@iu.edu.
KEYWORDS
Funding information
No funding was received for this study. genital infection, network meta-analysis, SGLT2 inhibitor, type 2 diabetes, urinary tract
infection

1 | I N T RO D UC T I O N either insulin secretion or insulin sensitivity are always associated

with gradual loss of efficacy over time owing to the progressive loss
Type 2 diabetes mellitus (T2DM) is a complex progressive metabolic of β-cell function5 and to metabolic comorbidities such as obesity,
disorder associated with macrovascular and microvascular hypoglycaemia and heart failure.6,7
complications,1 which has already become the leading cause of death Sodium-glucose co-transporter 2 (SGLT2) inhibitors are a novel
1,2
and disability worldwide. In addition to lifestyle management, oral class of OADs for treating T2DM. With an action independent of
antidiabetic drugs (OADs), including metformin, sulphonylureas and insulin, they can significantly reduce glycated haemoglobin (HbA1c)
thiazolidinediones, are of great importance for the pharmacological levels in drug-naive patients8 or those inadequately controlled by
treatment of T2DM;3,4 however, therapeutic strategies that improve metformin.9 In addition, SGLT2 inhibitors exert multiple metabolic

348 © 2016 John Wiley & Sons Ltd wileyonlinelibrary.com/journal/dom Diabetes Obes Metab 2017; 19(3):348–355
LI ET AL.
benefits including weight loss, blood pressure reduction and increase
10–12
in HDL cholesterol level. Furthermore, SGLT2 inhibitors might
also provide cardiovascular protections.13 An increased incidence of
urinary tract infections (UTIs) and genital infections, however, was
observed in patients treated with SGLT2 inhibitors including
14 15 16
dapagliflozin, canagliflozin and empagliflozin.
Furthermore, in December 2015, the US Food and Drug Admin-
istration added a warning for SGLT2 inhibitors about the potential
serious side effect of UTIs.17
Meta-analysis based on published trials showed a higher risk of
UTIs and genital infections in patients treated with SGLT2 inhibitors
compared with placebo,18 but did not discuss the possible discrep-
ancy relating to their structural differences. In addition, high risk
factors that contributed to these infectious diseases have been
discussed,19 but never quantitatively assessed. We therefore con-
ducted a pairwise meta-analysis and network meta-analysis to
explore the safety of SGLT2 inhibitors with regard to the genital and
urinary tracts, and examined possible risk factors.
2 | MATERIALS AND METHODS
2.1 | Search strategy
Electronic databases including PubMed, Embase, Cochrane Central
Register of Controlled Trials (CENTRAL) and ClinicalTrials.gov were
systematically searched to identify randomized controlled trials
(RCTs) published up to October 9, 2016 in patients with T2DM trea-
ted with SGLT2 inhibitors. The search terms included “sodium-
glucose cotransporter”, “SGLT2”, “SGLT-2”, individual names of
SGLT-2 inhibitors, “random” and “RCTs” (Table S1). Furthermore,
we searched for additional eligible trials from the reference lists
of retrieved meta-analyses, review articles and other relevant
publications.
2.2 | Study selection and data extraction
Two reviewers independently selected the citations retrieved from
electronic databases according to the following inclusion criteria:
(1) RCTs with full-text publication; (2) follow-up duration ≥24 weeks;
(3) comparison of SGLT2 inhibitors with placebo or other OADs in
adults with T2DM; and (4) RCTs that reported UTIs or/and genital
infections as adverse events (or serious adverse events) according to
prespecified lists from the Medical Dictionary for Regulatory Activ-
ities (MedDRA). Two reviewers independently screened all citations
obtained from the initial search. Any discrepancy was referred to a
third reviewer and resolved by discussion.
Two reviewers independently extracted data with a self-designed
standardized data extraction form including: first author (publication
year); study design (monotherapy or add-on to background treat-
ments); country of origin; length of follow-up; sample size; patient
characteristics (eg, inclusion criteria, background treatments, ethnic-
ity, mean age, proportion of men and duration of T2DM); control
treatments; interventions (type and dose of SGLT2 inhibitor); and
outcomes of interest.
349
Primary outcomes included incidence of UTIs and genital infec-
tions. Whenever possible, data obtained from the longest follow-up
were extracted from publications or ClinicalTrials.gov. For RCTs
which did not report outcomes of UTIs or genital infections, sugges-
tive events or consistent events judged by signs and subjective com-
plains were deemed equivalent and extracted. Furthermore, the
occurrence of different types of these outcomes was further explored
using the meta-regression method.
As we focused on the effects of different SGLT2 inhibitors, other
active treatments including metformin, glimepiride, sitagliptin, saxa-
gliptin and linagliptin were referred to as an active treatment arm.
For trials with several arms, the arms with combination therapy (eg,
saxagliptin plus dapagliflozin) or not commonly used doses (200 mg
canagliflozin or 1 mg dapagliflozin) were excluded. Publications that
contained two trials comparing groups of non-overlapping patients
with different background treatments (eg, monotherapy or add-on
therapy) were considered separately.8,9,20 In cases of missing or
unclear data, the original authors were contacted for more
information.
2.3 | Risk of bias
Two reviewers assessed the quality of included RCTs according to
the Cochrane risk-of-bias tool.21 The quality of trials was judged as
low, unclear or high risk of bias based on the following domains: ran-
dom sequence generation (selection bias); allocation concealment
(selection bias); blinding (performance bias and detection bias);
incomplete outcome data (attrition bias); and selective reporting
(reporting bias).
2.4 | Statistical analysis
Data management, transformation and calculation were performed
using STATA 13.1 in the pairwise and network meta-analyses.22 Odds
ratios (ORs) and 95% confidence intervals (CIs) were used to calculate
the comparative effect sizes.
First, we performed standard pairwise meta-analysis with random
effect models.23 Heterogeneity between studies was assessed and
judged as low (<25%), moderate (25%-75%) and high (>75%) using
the I2 statistic. Second, network meta-analysis was performed using
random effect models to preserve comparisons of randomized treat-
ments within trials assuming a common heterogeneity variable for
each network (the tau [τ] value).24 Heterogeneity was deemed to be
reasonable (0.1-0.5), high (0.5-1.0) or extremely high (>1.0) according
to the τ value.25 Relative ranking probabilities of a specified outcome
were estimated by surface under the cumulative ranking curve
(SUCRA) probabilities and mean ranks. A 0.5 zero-cell correction was
applied before meta-analysis whenever zero events occurred.26
Consistency within every closed triangular or quadratic loop was
assessed using the “loop-specific” approach, and its 95% CI, excluding
zero, was judged to be inconsistent.27,28 In addition, we checked the
global heterogeneity in networks by the “design-by-treatment” inter-
action model.29 Publication bias was tested by visual inspection of
funnel plots.30,31 Furthermore, random-effect network meta-
regression analyses were conducted to assess baseline trial
350
characteristics, including study design (mode), trial duration, sample
size, percentage of men, age, baseline HbA1c, body mass index and
duration of T2DM, as effect modifiers on estimates for primary
outcomes.
3 | RESULTS
Of 209 potentially eligible trials, 52 RCTs involving 36 689 adults
were included in this network meta-analysis (Table S2). A flow chart
showing the identification of eligible trials is provided in Figure 1. A
total of 11 trials examined different doses for canagliflozin, 21 trials
assessed dapagliflozin, 18 trials evaluated empagliflozin, and 1 trial
each evaluated luseogliflozin and ipragliflozin. Treatment arms that
combined SGLT2 inhibitors with other active treatments in 7 studies
were excluded.8,9,32–36
The follow-up periods of included trials ranged from 24 to
160 weeks. The number of patients in each study ranged from
LI ET AL.
145 to 7020, and the baseline HbA1c ranged from 7.2% to 9.3%.
UTIs, genital infections and consistent events were reported in each
trial according to MedDRA preferred terms, while UTIs and genital
infections were diagnosed based on signs, symptoms and other
reports suggesting infection events in the urinary or genital tract at
each study visit.
Overall, the risk of bias was high or unclear because of random
sequence generation in 12 trials (23.1%), allocation concealment in
22 trials (42.3%), blinding of participants and personnel in 16 trials
(30.8%), blinding of assessor in 32 trials (61.5%) and selective report-
ing in 30 trials (57.7%; Figure S1). There was a high percentage of
selective reporting because trials that reported consistent or sugges-
tive events were deemed to be high risk.
Networks of eligible comparisons for the primary outcomes are
shown in Figure 2. There was no inconsistency between evidence
derived from direct and indirect comparisons in any loop because
their 95% CIs all included zero (Table S3). The design-by-treatment
interaction model did not detect global inconsistency in any
FIGURE 1 Flow chart of the identification
of eligible trials.
LI ET AL. 351

FIGURE 2 Network plot of available

treatments for category effect (A) and
dose effect (B) of SGLT2 inhibitors on
primary outcomes. ACT, active
treatments; CANA, canagliflozin; DAPA,
dapagliflozin; EMPA, empagliflozin; IPRA,
ipragliflozin; LUSEO, luseogliflozin; PLA,
placebo.

treatment networks (P all > .05). In addition, no evidence of publica-
tion bias in the outcomes of UTIs or genital infection was observed
from the visual inspection of funnel plots (Figure S2). Contributions
of direct evidence to network analyses are shown in Table S4. Fur-
thermore, the comparative effects of all treatments against placebo
were ranked according to SUCRA probabilities (Table S5).
A total of 50 trials involving 35 060 patients were available for
category analysis of SGLT2 inhibitors on risk of UTIs and 34 728
patients from 49 trials for genital infection. Among these SGLT2 inhi-
bitors, only dapagliflozin was associated with significantly higher UTI
risk than placebo (OR 1.28, 95% CI 1.06-1.54; Table 1) or other
active treatments (OR 1.28, 95% CI 1.02-1.61). Empagliflozin was
associated with a significantly lower risk of UTIs as compared with
dapagliflozin (OR 0.79, 95% CI 0.64-0.97; Table 1). Meanwhile, empa-
gliflozin, dapagliflozin and canagliflozin were associated with a higher
risk of genital infections than placebo: ORs were 3.64 (95% CI 2.87-
4.63), 4.51 (95% CI 3.37-6.04) and 4.99 (95% CI 3.74-6.67), respec-
tively (Table 1).
A total of 51 trials involving 35 875 adults provided adequate
data for analysis of different doses of SGLT2 inhibitors on risk of
UTIs and genital infections. Dapagliflozin 10 mg was associated with
a significantly increased risk of UTI as compared with placebo
(OR 1.32, 95% CI 1.08-1.62), empagliflozin 25 mg (OR 1.33, 95% CI
1.06-1.67), and other active treatments (OR 1.30, 95% CI 1.00-1.67;
Table 2). In addition, dapagliflozin 10 mg was associated with more
UTI events than dapagliflozin 2.5 mg (OR 1.65, 95% CI 1.05-2.57;
Table 2).
A total of 1877 genital infection events contributed to the dose
effect analysis, most of which were vulvovaginitis (infection of the
vulva and vagina) in women and balanitis in men.37–39 All doses of
canagliflozin, dapagliflozin and empagliflozin were associated with
more genital infections than placebo, and the ORs ranged from 3.21
(95% CI 2.08-4.93) for dapagliflozin 2.5 mg to 5.23 (95% CI 3.86-
7.09) for canagliflozin 300 mg (Table 2). Furthermore, dapagliflozin
10 mg was associated with significantly higher risk of genital infec-
tions than dapagliflozin 2.5 mg (OR 1.55, 95% CI 1.08-2.23; Table 2).
Meanwhile, all doses of canagliflozin, dapagliflozin and empagaliflozin
were associated with a higher risk of genital infections than other
active treatments and no significant differences were detected
among other comparisons (Table 2).
Similarly to the results from network meta-analysis, results from
pairwise meta-analysis are shown in Table S6.
Subgroup analysis of gender effect on risk of UTIs was feasible
for empagliflozin in 4 trials and for dapagliflozin in 8 trials, and results
from the pairwise meta-analysis showed that occurrence of UTIs was
significantly higher in women than men for empagliflozin (OR 4.92,
95% CI 4.22-5.74; P = .082) and dapagliflozin (OR 3.08, 95% CI 2.08-
4.55; P = .664). At the same time, more UTI events in women were
also observed in the placebo group (OR 6.24, 95% CI 5.08-7.66;
P = .545) without heterogeneity (I2 = 0%). Similarly, there was a
higher incidence of genital infections in women for all SGLT2 inhibi-
tors (OR 2.25, 95% CI 1.99-2.54) and placebo (OR 2.48, 95% CI 1.66-
3.73) when compared with men; however, gender was not a factor
that had a significant modification effect on UTIs or genital infections
based on meta-regression analysis (Table S7).
Meta-regression analyses were also performed to assess the
characteristics of studies (eg, study mode, follow-up periods, type of
events and sample size) as well as those of participants (eg, race,

TABLE 1 Network meta-analysis results of category effect of SGLT2 inhibitors on UTI (lower triangle) and genital infections (upper triangle)

ACT 0.98 (0.02, 50.43) 0.58 (0.02, 14.76) 0.26 (0.19, 0.34) 0.21 (0.15, 0.29) 0.19 (0.14, 0.26) 0.93 (0.70, 1.25)
0.33 (0.06, 1.72) IPRA 0.60 (0.00, 95.79) 0.26 (0.01, 13.45) 0.21 (0.00, 10.90) 0.19 (0.00, 9.84) 0.95 (0.02, 48.65)
0.63 (0.02, 15.72) 1.87 (0.05, 69.02) LUSEO 0.44 (0.02, 11.10) 0.36 (0.01, 9.00) 0.32 (0.01, 8.13) 1.60 (0.06, 40.08)
0.99 (0.84, 1.18) 2.97 (0.58, 15.20) 1.59 (0.06, 39.73) EMPA 0.81 (0.57, 1.14) 0.73 (0.52, 1.03) 3.64 (2.87, 4.63)
0.78 (0.62, 0.98) 2.34 (0.45, 12.07) 1.25 (0.05, 31.45) 0.79 (0.64, 0.97) DAPA 0.90 (0.62, 1.32) 4.51 (3.37, 6.04)
0.88 (0.71, 1.08) 2.62 (0.51, 13.53) 1.40 (0.06, 35.23) 0.88 (0.72, 1.08) 1.12 (0.87, 1.45) CANA 4.99 (3.74, 6.67)
1.00 (0.84, 1.20) 3.00 (0.59, 15.31) 1.60 (0.06, 40.08) 1.01 (0.91, 1.12) 1.28 (1.06, 1.54) 1.14 (0.94, 1.39) PLA

Comparisons between treatments should be read from left to right and the estimate shown as OR with 95% CI is in the common cell between the
column-defining treatment and the row-defining treatment. ORs < 1 favour the column-defining treatment. Statistically different results are in bold and
underscored.
ACT, active treatments; CANA, canagliflozin; DAPA, dapagliflozin; EMPA, empagliflozin; IPRA, ipragliflozin; LUSEO, luseogliflozin; PLA, placebo.
 352 LI ET AL.

0.60 (0.00, 95.79) 0.26 (0.01, 13.27) 0.25 (0.00, 12.98) 0.19 (0.00, 9.92) 0.22 (0.00, 11.47) 0.30 (0.01, 15.53) 0.18 (0.00, 9.40) 0.20 (0.00, 10.09) 0.95 (0.02, 48.65)
1.60 (0.06, 40.08)

Comparisons between treatments should be read from left to right and the estimate shown as OR with 95% CI is in the common cell between the column-defining treatment and the row-defining treatment. ORs < 1
baseline HbA1c level, T2DM duration, body mass index and age) on

3.70 (2.87, 4.76)

3.78 (2.92, 4.89)
4.97 (3.60, 6.86)
4.31 (2.99, 6.22)
3.21 (2.08, 4.93)
5.23 (3.86, 7.09)
4.88 (3.59, 6.64)
0.96 (0.71, 1.31) effect sizes of outcomes, only follow-up periods had a marginal modi-
fication effect on the risk of UTIs for canagliflozin (Table S7).

PLA
4 | DI SCU SSION
0.18 (0.13, 0.26) 0.20 (0.14, 0.28)

0.50 (0.02, 12.86) 0.31 (0.01, 7.77) 0.33 (0.01, 8.33)

0.71 (0.49, 1.01) 0.76 (0.53, 1.09)
0.72 (0.50, 1.04) 0.77 (0.54, 1.12)
0.95 (0.62, 1.44) 1.02 (0.67, 1.56)
0.82 (0.52, 1.30) 0.88 (0.56, 1.40)
0.61 (0.37, 1.02) 0.66 (0.39, 1.10)
1.07 (0.89, 1.29)

1.14 (0.92, 1.42) 1.19 (0.96, 1.48)

0.96 (0.78, 1.18) CANA100
Our meta-analysis results indicate that all doses of canagliflozin,
dapagliflozin and empagliflozin were associated with a significantly
higher risk of genital infections than placebo and other active treat-
ments, but only dapagliflozin 10 mg led to a higher risk of UTIs than
placebo and other active treatments. Furthermore, dapagliflozin
CANA300

appeared to have a dose–response relationship for risk of UTIs and

genital infections. Higher doses of dapagliflozin were associated with
more UTI and genital infection events than lower doses of dapagliflo-
0.30 (0.18, 0.49)

1.55 (1.08, 2.23)

1.15 (0.71, 1.87)
1.18 (0.72, 1.92)

1.34 (0.92, 1.96)

0.70 (0.43, 1.16)

0.67 (0.41, 1.11)
0.80 (0.51, 1.26)

zin according to the ORs and SUCRA ranking. By contrast, incidence

of UTIs and genital infections remained similar or even decreased for
DAPA2.5

empagliflozin and canagliflozin. No significant differences were

Network meta-analysis results of dose effect of SGLT2 inhibitors on UTI (lower triangle) and genital infections (upper triangle)

detected between lesuogliflozin or ipragliflozin and placebo or active

ACT, active treatments; CANA, canagliflozin; DAPA, dapagliflozin; EMPA, empagliflozin; IPRA, ipragliflozin; LUSEO, luseogliflozin; PLA, placebo.

treatments; however, conclusions should be drawn with caution

0.19 (0.13, 0.29) 0.22 (0.15, 0.34)

0.43 (0.02, 10.95) 0.42 (0.02, 10.72) 0.32 (0.01, 8.20) 0.37 (0.01, 9.49)
0.74 (0.50, 1.10) 0.86 (0.56, 1.31)
0.76 (0.51, 1.13) 0.88 (0.57, 1.35)
1.15 (0.86, 1.54)

1.65 (1.05, 2.57) 1.45 (0.91, 2.31)

1.15 (0.86, 1.54) 1.02 (0.71, 1.46)
1.11 (0.83, 1.49) 0.98 (0.68, 1.41)

1.32 (1.08, 1.62) 1.16 (0.86, 1.57)

because of sparse data (only one RCT was included for each drug).
Factors that affect these infectious events have been discussed
previously, but no definite conclusion was drawn. Hyperglycaemia
1.14 (0.84, 1.53) DAPA5

has been speculated to increase the genital infection and UTI risk
because it is known to weaken the immune system40 and cause
infections, including oral thrush, skin furunculosis and panniculitis.41
However, it came into conflict with the fact that SGLT2 inhibitors
DAPA10

could significantly reduce HbA1c,42 but led to a higher risk of infec-

tious events than placebo. The relationship between baseline HbA1c
level and UTIs was also explored in our meta-regression analysis, and
0.25 (0.19, 0.35)

0.98 (0.83, 1.16)

0.80 (0.64, 1.01)

0.91 (0.66, 1.26)
1.33 (0.83, 2.12)
0.93 (0.73, 1.18)
0.89 (0.70, 1.14)
1.06 (0.94, 1.20)

no statistical difference was detected. Conversely, glucosuria was

favour the column-defining treatment. Statistically different results are in bold and underscored.

considered to be a more likely risk factor because SGLT2 inhibitors

EMPA10

showed their anti-hyperglycaemic effects by inhibiting glucose reab-

sorption from the kidney, and thus resulting in high glucosuria.39,41
Pharmacodynamics research indicated a dose-related increase in the
1.01 (0.02, 52.03) 0.60 (0.02, 15.23) 0.26 (0.19, 0.35)

2.27 (0.44, 11.74) 1.21 (0.05, 30.55) 0.75 (0.60, 0.94)

2.82 (0.55, 14.46) 1.51 (0.06, 37.79) 0.93 (0.83, 1.04)

2.58 (0.49, 13.54) 1.38 (0.05, 34.96) 0.85 (0.62, 1.17)

3.74 (0.69, 20.31) 2.00 (0.08, 51.59) 1.23 (0.77, 1.96)
2.62 (0.51, 13.57) 1.40 (0.06, 35.29) 0.86 (0.68, 1.09)
2.52 (0.49, 13.05) 1.35 (0.05, 33.92) 0.83 (0.65, 1.05)
3.00 (0.59, 15.31) 1.60 (0.06, 40.08) 0.99 (0.88, 1.11)

cumulative amount of glucose excreted in the urine over 24 hours

for dapagliflozin,43 but not for canagliflozin44 or empagliflozin.45 Fur-
3.04 (0.59, 15.57) 1.62 (0.06, 40.68) EMPA25

thermore, patients with familial renal glucosuria, a syndrome identi-

fied as glucosuria but with normal glycaemia, rarely experienced
UTIs;46 therefore, the contribution of glucosuria to the observed
increased risk of genital infections and UTIs in patients with T2DM
remained unclear.
1.87 (0.05, 69.02) LUSEO2.5

Previous studies showed that genital infections and UTIs associ-

ated with SGLT2 inhibitors were less frequent in men than in
women;39,47,48 however, our analysis showed that even though more
infectious events occurred in women than men for empagliflozin and
dapagliflozin, it should not be ascribed to a gender effect because a
IPRA50

similar tendency was observed in the placebo group. In fact, infec-

tious events were always more common in women even in those
without SGLT2 inhibitors.49,50 Furthermore, our meta-regression
0.64 (0.03, 16.01)

0.77 (0.60, 1.00)

0.34 (0.07, 1.75)

1.03 (0.86, 1.24)

0.96 (0.79, 1.16)

0.88 (0.63, 1.23)

1.27 (0.79, 2.06)
0.89 (0.71, 1.13)
0.86 (0.68, 1.09)
1.02 (0.85, 1.23)

analysis did not identify the significant gender effect between any
treatment group and placebo group. UTIs and genital infections were
mostly reported within the first 24 to 26 weeks of SGLT2 inhibitor
TABLE 2

ACT

therapy,51–53 and the occurrences decreased over time.54 This was

consistent with our finding that follow-up periods had no overall
LI ET AL.
modification effect on the primary outcomes and suggested the long-
term safety of SGLT2 inhibitors for UTIs and genital infections.
The present analysis has several advantages. First of all, unlike
two previous meta-analyses or reviews,18,19 we had a comprehensive
strategy to systematically search and retrieve potentially eligible
RCTs, which avoided selection bias. Second, the rigorous inclusion
criteria for our meta-analysis helped to increase the internal validity
of our results. Third, compared with traditional pairwise meta-analy-
sis, our network meta-analysis was able to integrate both direct and
indirect comparison results by synthesizing all available evidence on
UTIs and genital infections; we could therefore obtain comparison
results when no direct head-to-head comparisons were available,
such as empagliflozin 25 mg vs dapagliflozin 10 mg. Fourth, we eval-
uated the effect of dose on these infectious events, which was over-
looked by two previous meta-analyses.18,55
Compared with a recent network-meta analysis performed by
Zaccardi et al,42 which included 38 trials published before November
2015, the literature search in the present study was updated to
October 2016 and we additionally included 15 trials and two new
SGLT2 inhibitors (luseogliflozin 2.5 mg and ipragliflozin 50 mg). Fur-
thermore, our meta-analysis showed that dapagliflozin was associated
with an increased risk of urinary and genital infections, possibly in a
dose–response relationship, which was not the case for canagliflozin
or empagliflozin. Finally, to our knowledge, this was the first study to
quantitatively explore the possible risk factors that might lead to risk
of UTIs and genital infections. Gender and baseline HbA1c levels
which were investigated previously,19 as well as other possible fac-
tors (eg, ethnicity, body mass index and T2DM duration) were
assessed by our meta-regression analysis.
We also acknowledge the following study limitations. First, the
judgment of suggestive events was based on signs and subjective
complaints, while consistent events were reported according to pro-
spectively self-defined preferred items, which varied among differ-
ent researchers. To simplify the process of analysis, we pooled
these events instead of analysing them separately. Such risk of bias
might be higher for genital infections because significant differences
were detected for dapagliflozin with regard to type of genital infec-
tions. Second, if results from multiple follow-ups were found for a
trial, results from the longest follow-up were used in our meta-anal-
ysis. This might have introduced bias because the meta-regression
analysis indicated that follow-up period might modify the effect of
canagliflozin on category effects of UTIs; however, significant modi-
fication effects of follow-up duration and outcome type from spe-
cific SGLT2 inhibitors might be attributable to the combination of
different doses as no significant effects were found for dose effects
of primary outcomes.
In conclusion, the present study shows that all SGLT2 inhibitors
were associated with a significantly higher number of genital infec-
tions but only dapagliflozin 10 mg led to significant risk of UTIs.
Dapagliflozin appeared to have a dose–response relationship for
both outcomes. No prespecified factors were identified to signifi-
cantly affect the risk of UTIs or genital infections for SGLT2 inhibi-
tors; therefore, UTIs and genital infections should be monitored
cautiously for all patients, even those deemed to be at lower risk
such as men.
353
Conflict of interest
The authors declared no conflict of interest.
Author contributions
D. L. and H. T. had the idea for the study and led the study design. D.
L., T. W., S. S., Z. F., Y. D., and H. T. identified and selected trials and
extracted data. D. L., S. S., and H. T. performed all data analyses,
checked for statistical consistency and interpreted the results. D. L.,
T. W., S. S., Z. F., Y. D., and H. T. contributed to data interpretation.
D. L., T. W., and H. T. drafted the report, and all other authors (S. S.,
Z. F., and Y. D.) critically reviewed the report.
RE FE RE NC ES
1. Grundy SM. Metabolic syndrome pandemic. Arterioscler Thromb Vasc
Biol. 2008;28:629-636.
2. Lozano R, Naghavi M, Foreman K, et al. Global and regional mortality
from 235 causes of death for 20 age groups in 1990 and 2010: a sys-
tematic analysis for the Global Burden of Disease Study 2010. Lancet.
2012;380:2095-2128.
3. Kahn SE, Haffner SM, Heise MA, et al. Glycemic durability of rosiglita-
zone, metformin, or glyburide monotherapy. N Engl J Med. 2006;
355:2427-2443.
4. Nathan DM, Buse JB, Davidson MB, et al. Management of hypergly-
cemia in type 2 diabetes: a consensus algorithm for the initiation and
adjustment of therapy: a consensus statement from the American
Diabetes Association and the European Association for the Study of
Diabetes. Diabetes Care. 2006;29:1963-1972.
5. Inzucchi SE, Bergenstal RM, Buse JB, et al. Management of hypergly-
caemia in type 2 diabetes: a patient-centered approach. Position
statement of the American Diabetes Association (ADA) and the
European Association for the Study of Diabetes (EASD). Diabetologia.
2012;55:1577-1596.
6. Brunton SA. Hypoglycemic potential of current and emerging pharma-
cotherapies in type 2 diabetes mellitus. Postgrad Med. 2012;124:
74-83.
7. Aguilar RB. Evaluating treatment algorithms for the management of
patients with type 2 diabetes mellitus: a perspective on the definition
of treatment success. Clin Ther. 2011;33:408-424.
8. Lewin A, DeFronzo RA, Patel S, et al. Initial combination of empagli-
flozin and linagliptin in subjects with type 2 diabetes. Diabetes Care.
2015;38:394-402.
9. DeFronzo RA, Lewin A, Patel S, et al. Combination of empagliflozin and
linagliptin as second-line therapy in subjects with type 2 diabetes inad-
equately controlled on metformin. Diabetes Care. 2015;38:384-393.
10. Stenlof K, Cefalu WT, Kim KA, et al. Efficacy and safety of canagliflo-
zin monotherapy in subjects with type 2 diabetes mellitus inade-
quately controlled with diet and exercise. Diabetes Obes Metab.
2013;15:372-382.
11. Wilding JP, Charpentier G, Hollander P, et al. Efficacy and safety of
canagliflozin in patients with type 2 diabetes mellitus inadequately
controlled with metformin and sulphonylurea: a randomised trial. Int J
Clin Pract. 2013;67:1267-1282.
12. Rosenstock J, Jelaska A, Frappin G, et al. Improved glucose control
with weight loss, lower insulin doses, and no increased hypoglycemia
with empagliflozin added to titrated multiple daily injections of insulin
in obese inadequately controlled type 2 diabetes. Diabetes Care.
2014;37:1815-1823.
13. Zinman B, Wanner C, Lachin JM, et al. Empagliflozin, cardiovascular
outcomes, and mortality in type 2 diabetes. N Engl J Med.
2015;373:2117-2128.
14. Jabbour SA, Hardy E, Sugg J, Parikh S. Dapagliflozin is effective as
add-on therapy to sitagliptin with or withoutmetformin: a 24-week,
multicenter, randomized, double-blind, placebo-controlled study.
Diabetes Care. 2014;37:740-750.
354
15. Bode B, Stenlof K, Harris S, et al. Long-term efficacy and safety of
canagliflozin over 104 weeks in patients aged 55-80 years with type
2 diabetes. Diabetes Obes Metab. 2015;17:294-303.
16. Haering HU, Merker L, Christiansen AV, et al. Empagliflozin as add-on
to metformin plus sulphonylurea in patients with type 2 diabetes.
Diabetes Res Clin Pract. 2015;110:82-90.
17. US Food and Drug Administration. FDA Drug Safety Communication:
FDA revises labels of SGLT2 inhibitors for diabetes to include warn-
ings about too much acid in the blood and serious urinary tract infec-
tions. 2015. http://www.fda.gov/Drugs/DrugSafety/ucm475463.htm.
Accessed May 15, 2015.
18. Wu JH, Foote C, Blomster J, et al. Effects of sodium-glucose
cotransporter-2 inhibitors on cardiovascular events, death, and major
safety outcomes in adults with type 2 diabetes: a systematic review
and meta-analysis. Lancet Diabetes Endocrinol. 2016;4:411-419.
19. Arakaki RF. Sodium-glucose cotransporter-2 inhibitors and genital
and urinary tract infections in type 2 diabetes. Postgrad Med.
2016;128:409-417.
20. Ferrannini E, Berk A, Hantel S, et al. Long-term safety and efficacy of
empagliflozin, sitagliptin, and metformin: an active-controlled,
parallel-group, randomized, 78-week open-label extension study in
patients with type 2 diabetes. Diabetes Care. 2013;36:4015-4021.
21. Higgins JPT. ADSA. Chapter 8: assessing risk of bias in included stud-
ies. In: Higgins J, Green S, eds. Cochrane Handbook for Systematic
Reviews of Interventions Version 510 (updated March 2011). The
Cochrane Collaboration; 2011. http://handbook.cochrane.org/
chapter_8/8_assessing_risk_of_bias_in_included_studies.htm.
Accessed March 25, 2016.
22. White IR, Barrett JK, Jackson D, Higgins JP. Consistency and incon-
sistency in network meta-analysis: model estimation using multivari-
ate meta-regression. Res Synth Methods. 2012;3:111-125.
23. Bucher HC, Guyatt GH, Griffith LE, Walter SD. The results of direct
and indirect treatment comparisons in meta-analysis of randomized
controlled trials. J Clin Epidemiol. 1997;50:683-691.
24. Salanti G, Marinho V, Higgins JP. A case study of multiple-treatments
meta-analysis demonstrates that covariates should be considered.
J Clin Epidemiol. 2009;62:857-864.
25. Turner RM, Davey J, Clarke MJ, Thompson SG, Higgins JP. Predicting
the extent of heterogeneity in meta-analysis, using empirical data
from the Cochrane Database of Systematic Reviews. Int J Epidemiol.
2012;41:818-827.
26. Keus F, Wetterslev J, Gluud C, Gooszen HG, van Laarhoven CJ.
Robustness assessments are needed to reduce bias in meta-analyses
that include zero-event randomized trials. Am J Gastroenterol.
2009;104:546-551.
27. Estacio RO, Coll JR, Tran ZV, Schrier RW. Effect of intensive blood
pressure control with valsartan on urinary albumin excretion in nor-
motensive patients with type 2 diabetes. Am J Hypertens.
2006;19:1241-1248.
28. Abe H, Minatoguchi S, Ohashi H, et al. Renoprotective effect of the
addition of losartan to ongoing treatment with an angiotensin con-
verting enzyme inhibitor in type-2 diabetic patients with nephropa-
thy. Hypertens Res. 2007;30:929-935.
29. Higgins JP, Jackson D, Barrett JK, Lu G, Ades AE, White IR. Consist-
ency and inconsistency in network meta-analysis: concepts and mod-
els for multi-arm studies. Res Synth Methods. 2012;3:98-110.
30. Sterne JA, Sutton AJ, Ioannidis JP, et al. Recommendations for exam-
ining and interpreting funnel plot asymmetry in meta-analyses of ran-
domised controlled trials. BMJ. 2011;343:d4002.
31. Sterne JA, Egger M. Funnel plots for detecting bias in meta-analysis:
guidelines on choice of axis. J Clin Epidemiol. 2001;54:1046-1055.
32. Rosenstock J, Hansen L, Zee P, et al. Dual add-on therapy in type
2 diabetes poorly controlled with metformin monotherapy: a rando-
mized double-blind trial of saxagliptin plus dapagliflozin addition ver-
sus single addition of saxagliptin or dapagliflozin to metformin.
Diabetes Care. 2015;38:376-383.
33. Henry RR, Murray AV, Marmolejo MH, Hennicken D, Ptaszynska A,
List JF. Dapagliflozin, metformin XR, or both: initial pharmacotherapy
for type 2 diabetes, a randomised controlled trial. Int J Clin Pract.
2012;66:446-456.
LI ET AL.
34. Frias JP, Guja C, Hardy E, et al. Exenatide once weekly plus dapagli-
flozin once daily versus exenatide or dapagliflozin alone in patients
with type 2 diabetes inadequately controlled with metformin mono-
therapy (DURATION-8): a 28 week, multicentre, double-blind, phase
3, randomised controlled trial. Lancet Diabetes Endocrinol.
2016;4:1004-1016.
35. Hadjadj S, Rosenstock J, Meinicke T, Woerle HJ, Broedl UC. Initial
combination of empagliflozin and metformin in patients with type
2 diabetes. Diabetes Care. 2016;39:1718-1728.
36. Rosenstock J, Chuck L, Gonzalez-Ortiz M, et al. Initial combination
therapy with canagliflozin plus metformin versus each component as
monotherapy for drug-naive type 2 diabetes. Diabetes Care.
2016;39:353-362.
37. Lavalle-Gonzalez FJ, Januszewicz A, Davidson J, et al. Efficacy and
safety of canagliflozin compared with placebo and sitagliptin in
patients with type 2 diabetes on background metformin monother-
apy: a randomised trial. Diabetologia. 2013;56:2582-2592.
38. Nyirjesy P, Zhao Y, Ways K, Usiskin K. Evaluation of vulvovaginal
symptoms and Candida colonization in women with type 2 diabetes
mellitus treated with canagliflozin, a sodium glucose co-transporter
2 inhibitor. Curr Med Res Opin. 2012;28:1173-1178.
39. Johnsson KM, Ptaszynska A, Schmitz B, Sugg J, Parikh SJ, List JF. Vul-
vovaginitis and balanitis in patients with diabetes treated with dapa-
gliflozin. J Diabetes Complications. 2013;27:479-484.
40. Geerlings SE, Hoepelman AI. Immune dysfunction in patients with
diabetes mellitus (DM). FEMS Immunol Med Microbiol. 1999;26:
259-265.
41. Casqueiro J, Casqueiro J, Alves C. Infections in patients with diabetes
mellitus: a review of pathogenesis. Indian J Endocrinol Metab.
2012;16(suppl 1):S27-S36.
42. Zaccardi F, Webb DR, Htike ZZ, Youssef D, Khunti K, Davies MJ. Effi-
cacy and safety of sodium-glucose cotransporter 2 inhibitors in type
2 diabetes mellitus: systematic review and network meta-analysis.
Diabetes Obes Metab. 2016;18:783-789.
43. Kasichayanula S, Chang M, Hasegawa M, et al. Pharmacokinetics and
pharmacodynamics of dapagliflozin, a novel selective inhibitor of
sodium-glucose co-transporter type 2, in Japanese subjects without
and with type 2 diabetes mellitus. Diabetes Obes Metab.
2011;13:357-365.
44. Devineni D, Curtin CR, Polidori D, et al. Pharmacokinetics and phar-
macodynamics of canagliflozin, a sodium glucose co-transporter
2 inhibitor, in subjects with type 2 diabetes mellitus. J Clin Pharmacol.
2013;53:601-610.
45. Zhao X, Cui Y, Zhao S, et al. Pharmacokinetic and pharmacodynamic
properties and tolerability of single- and multiple-dose once-daily
empagliflozin, a sodium glucose cotransporter 2 inhibitor, in Chinese
patients with type 2 diabetes mellitus. Clin Ther. 2015;37:1493-
1502.
46. Santer R, Calado J. Familial renal glucosuria and SGLT2: from a men-
delian trait to a therapeutic target. Clin J Am Soc Nephrol.
2010;5:133-141.
47. Schernthaner G, Gross JL, Rosenstock J, et al. Canagliflozin compared
with sitagliptin for patients with type 2 diabetes who do not
have adequate glycemic control with metformin plus sulfonylurea: a
52-week randomized trial. Diabetes Care. 2013;36:2508-2515.
48. Liakos A, Karagiannis T, Bekiari E, Boura P, Tsapas A. Update on long-
term efficacy and safety of dapagliflozin in patients with type 2 diabe-
tes mellitus. Ther Adv Endocrinol Metab. 2015;6:61-67.
49. Schneeberger C, Kazemier BM, Geerlings SE. Asymptomatic bacteriu-
ria and urinary tract infections in special patient groups: women
with diabetes mellitus and pregnant women. Curr Opin Infect Dis.
2014;27:108-114.
50. Hirji I, Guo Z, Andersson SW, Hammar N, Gomez-Caminero A. Inci-
dence of urinary tract infection among patients with type 2 diabetes
in the UK General Practice Research Database (GPRD). J Diabetes
Complications. 2012;26:513-516.
51. Bailey CJ, Morales VE, Woo V, Tang W, Ptaszynska A, List JF. Effi-
cacy and safety of dapagliflozin monotherapy in people with type
2 diabetes: a randomized double-blind placebo-controlled 102-week
trial. Diabet Med. 2015;32:531-541.
LI ET AL. 355

52. Ridderstrale M, Andersen KR, Zeller C, Kim G, Woerle HJ, Broedl UC. S U P P O R T I N G I N F O R M A T I ON
Comparison of empagliflozin and glimepiride as add-on to metformin
in patients with type 2 diabetes: a 104-week randomised, active-con-
trolled, double-blind, phase 3 trial. Lancet Diabetes Endocrinol. Additional Supporting Information may be found online in the sup-
2014;2:691-700. porting information tab for this article.
53. Leiter LA, Yoon KH, Arias P, et al. Canagliflozin provides durable
glycemic improvements and body weight reduction over 104 weeks
versus glimepiride in patients with type 2 diabetes on metformin: How to cite this article: Li D, Wang T, Shen S, Fang Z,
a randomized, double-blind, phase 3 study. Diabetes Care.
Dong Y and Tang H. Urinary tract and genital infections in
2015;38:355-364.
54. Forst T, Guthrie R, Goldenberg R, et al. Efficacy and safety of canagli- patients with type 2 diabetes treated with sodium-glucose co-
flozin over 52 weeks in patients with type 2 diabetes on background transporter 2 inhibitors: A meta-analysis of randomized con-
metformin and pioglitazone. Diabetes Obes Metab. 2014;16:467-477.
trolled trials, Diabetes Obes Metab, 2017;19(3):348–355.
55. Vasilakou D, Karagiannis T, Athanasiadou E, et al. Sodium-glucose
cotransporter 2 inhibitors for type 2 diabetes: a systematic review
and meta-analysis. Ann Intern Med. 2013;159:262-274.