Summary of the Evidence

Symptom relief and esophagitis healing (Patients with erosive esophagitis):

• In 12 head-to-head trials, there was no difference between lansoprazole, omeprazole, pantoprazole, and rabeprazole on the outcome of complete symptom relief at 4 weeks. The only significant difference on this outcome was in the comparison of esomeprazole 40 mg to omeprazole 20 mg. The pooled risk difference in three trials was 10% (95% CI 6%–14%).
• Esomeprazole 40 mg was also compared to lansoprazole 30 mg and to pantoprazole 40 mg for complete symptom relief at 4 weeks with no significant differences.
• Time to relief of heartburn was similar for all PPIs in head-to-head trials, but the methods used to measure and report this outcome varied.
• There is good evidence that there is no comparative difference between omeprazole, lansoprazole, pantoprazole, and rabeprazole for healing of esophagitis or relief of GERD symptoms. Thirteen head-to-head trials, 20 trials of these PPIs versus an H2-RA, and three systematic reviews found these four PPIs to be equally effective.
• Esomeprazole 40mg had higher 4- and 8-week healing rates than omeprazole 20mg.
• Three trials compared esomeprazole 40 mg to lansoprazole 30 mg. The pooled healing rate of two trials reporting healing at 4 weeks was 5% higher for esomeprazole (NNT = 20). One of three studies found a significantly higher healing rate for esomeprazole at 8 weeks (NNT=33). Two others found healing rates equivalent at 8 weeks, and the pooled estimate from 3 studies was not significant.

Symptom relief (patients with non-erosive or empirically-treated GERD

• Three head-to-head trials in patients with endoscopy-negative GERD found no difference between esomeprazole 20 mg and omeprazole 20 mg, pantoprazole 20 mg, and rabeprazole 10 mg. These studies used different outcome measures.
• Limited indirect evidence from placebo- and active-controlled trials suggests similar efficacy for heartburn resolution and complete symptom relief for all five PPIs.

Prevention of relapse (patients with erosive esophagitis):

• For maintenance of healed esophagitis, there is good evidence that there is no comparative difference between omeprazole, lansoprazole, and rabeprazole. The longest-term study (over 5 years) is of omeprazole versus rabeprazole.
• Two 6-month studies found lower relapse rates for esomeprazole 20 mg compared with lansoprazole 15 mg or pantoprazole 20 mg daily.
• Pantoprazole was more effective than ranitidine in one 12-month study, and esomeprazole was more effective that ranitidine in another 6-month study.

Prevention of relapse (patients with non-erosive or empirically treated GERD):

• In a 6-month head-to-head trial of on-demand esomeprazole vs continuous lansoprazole 15 mg in patients with endoscopy negative GERD, more patients discontinued lansoprazole (for any reason), but discontinuations because of heartburn were not significantly different between treatment groups.
• On-demand rabeprazole 10 mg, on-demand esomeprazole 20 mg, and continuous omeprazole 10 mg were more effective than placebo in prevention of relapse of symptoms over 6 months in patients with endoscopy negative GERD.
• There are no head-to-head trials of on-demand treatment of one PPI versus on-demand treatment of another PPI.

Results by baseline severity:

• Among patients with moderate to severe esophagitis at baseline, esomeprazole 40 mg was more effective at healing esophagitis at 4 and 8 weeks than omeprazole 20 mg and lansoprazole 30 mg.
• The pooled risk difference in 3 studies of omeprazole 20 mg vs esomeprazole 40 mg was 16% at 4 weeks and 13% at 8 weeks (NNT=6 at 4 weeks, 8 at 8 weeks)
• The pooled risk difference in 2 studies of lansoprazole 30 mg vs esomeprazole 40 mg was 8% at 4 weeks and 9% at 8 weeks (NNT=13 at 4 weeks, 11 at 8 weeks).
• In one study, pantoprazole 40 mg had a higher healing rate at 8 weeks than esomeprazole 40 mg in patients with moderate (Grade C) esophagitis at baseline.
• Lansoprazole 30 mg and omeprazole 20 mg had equivalent healing rates in patients with moderate to severe esophagitis in two studies.

Evidence in Children

• There are no direct comparisons of PPIs for reflux esophagitis in children. A fair quality placebo-controlled trial in infants did not find omeprazole to be superior to placebo in controlling symptoms or acid-exposure time.

Detailed Assessment

Erosive Esophagitis

We identified 26 randomized controlled trials comparing two or more PPIs in patients with endoscopically-proven GERD (Evidence Table 1).^{5–7, 10–32} One is unpublished,⁷ and two publications are supplemented with additional data provided by the manufacturer.^{5, 6} Omeprazole was the comparator in most studies. No study of omeprazole in combination with sodium bicarbonate met inclusion criteria. The scales used to grade esophagitis in these studies are described in Appendix F. The comparisons made in head-to-head studies are shown in Table 1 (the number of comparisons adds to 28 because 2 studies compared 3 different PPIs).

Table 1

Number of head-to-head trials, short-term treatment erosive GERD.

Three studies^{5, 12, 29} met all criteria for internal validity, one was rated poor,²¹ and the rest were fair. (Details of quality ratings of included trials are listed in Evidence Table 2.) Pregnant and lactating women, and women of childbearing potential were excluded from all studies, and the majority of patients enrolled were male. No children (i.e., under age 18) were included in these studies.

Relief of Symptoms

Four head-to-head comparisons of PPIs measured symptom relief as a primary outcome, ^{10, 11, 13, 16} and 14 reported symptoms as a secondary outcome.^{5, 6, 12, 14, 15, 17, 21–26, 30, 32} Symptoms in these studies were assessed through patient diaries, investigator-elicited reports, or both.

Complete symptom resolution

Fourteen head-to-head trials reported the proportion of patients with complete resolution of symptoms at 4 weeks.^{5, 6, 10, 12–14, 16, 17, 20, 23, 24, 26, 27, 29} We performed a random effects meta-analysis of data from these studies to determine an estimate of the proportion who were symptom free at 4 weeks for each drug. Results are shown in Table 2 below. Proportions ranged from 65% to 77%, and 95% confidence intervals overlapped, indicating the drugs are similarly efficacious for complete resolution of symptoms at 4 weeks.

Table 2

Symptom resolution in head-to-head trials in patients with erosive GERD.

Figure 3 shows risk differences in rates of complete symptom resolution at 4 weeks in these trials.^{5, 6, 10, 12–14, 16, 17, 20, 23, 24, 26, 27, 29} In Table 3 we report the difference in complete symptom resolution for comparisons of esomeprazole to other PPIs. The pooled data on the comparison of esomeprazole 40 mg to omeprazole 20 mg significantly favored esomeprazole 40mg; for every 10 persons treated with esomeprazole 40 mg versus omeprazole 20mg, one additional patient would be symptom-free at four weeks in the esomeprazole group. The pooled data for esomeprazole 40mg versus either lansoprazole 30mg or pantoprazole 40mg did not indicate a significant difference between the drugs.

Figure 3

Rates of complete resolution of symptoms at 4 weeks in head-to-head trials of PPIs (risk difference, 95% CI)

Table 3

Symptom resolution at 4 weeks in trials of esomeprazole vs another PPI in erosive GERD.

A single study reported complete resolution of symptoms after 1 week of therapy,³² finding rabeprazole 20 mg daily superior to omeprazole 20 mg daily (27.9% vs 16.6%, P = 0.0013 as calculated from number randomized and using Chi Square analysis).

Time to Relief of Symptoms

Thirteen studies reported the time to resolution of symptoms (no heartburn). This measure was reported as the percentage of patients with the outcome after a given time point (e.g., 1 day, 7 days), the median number of days to resolution, or both. In one study this outcome is reported as the number of days needed for 50% and 75% of patients to achieve relief of symptoms.¹⁰

Another measure used was the time to sustained resolution of heartburn, defined as the first of 7 consecutive days without heartburn. This outcome was used only in studies funded by the maker of esomeprazole, so it is not possible to compare this outcome on studies funded by others.

Esomeprazole vs omeprazole

In four studies that compared esomeprazole 40mg to omeprazole 20mg, the median number of days to the first resolution of symptoms was similar, but the median number of days to sustained resolution of symptoms favored esomeprazole in the 2 studies reporting this measure (Table 4).^{6, 12, 16} More patients taking esomeprazole 40 mg reached first of resolution of symptoms by 1 day and day 7 based on absolute proportions, than those taking omeprazole 20mg. These findings were statistically significant in one study,¹² non-significant in two others^{16, 31}, and not assessed in the fourth.⁶ The time to sustained resolution of heartburn was statistically superior with esomeprazole 40mg compared to omeprazole 20mg at 14 days in 2 studies.^{12, 16} The differences at other time points were mixed or not statistically assessed.

Table 4

Time to symptom relief in trials of esomeprazole vs. omeprazole in erosive GERD.

In a comparison of esomeprazole 20 mg to omeprazole 20 mg,⁶ a numerically higher proportion of omeprazole patients started 7 consecutive days without heartburn at day 1; esomeprazole had a higher proportion of patients with sustained relief by day 28; neither comparison was statistically significant, and the median number of days to sustained resolution was similar. This pattern was also seen in the time to first resolution of symptoms.

Esomeprazole vs lansoprazole

In three studies comparing esomeprazole 40 mg to lansoprazole 30 mg, results were mixed and outcomes were reported differently (Table 5). Overall, results did not favor one drug over another.

Table 5

Time to symptom relief in trials of esomeprazole vs. lansoprazole in erosive GERD.

Esomeprazole vs. pantoprazole

The two trials of esomeprazole versus pantoprazole reported these data differently and found conflicting results. In one trial of esomeprazole 40 mg versus pantoprazole 40 mg, more esomeprazole patients reached the start of sustained resolution of heartburn (7 consecutive days) after one day of treatment: 24% vs 20% (p-value not reported).³⁰ The median time to sustained resolution was 6 days vs 8 days (p<0.001). A second trial of esomeprazole 40 mg versus pantoprazole 40 mg compared the number of days it took for 50% and 75% of patients to achieve relief of heartburn.¹⁰ In both groups, 50% of patients had no heartburn after 2 days, but it took 3 days for 75% of the pantoprazole group to achieve relief of symptoms versus 8 days for the esomeprazole group. Confidence intervals for the number of days overlapped, however (2–7 days for pantoprazole vs. 3–14 days for esomeprazole).

Lansoprazole vs omeprazole

Three studies reported time to relief of heartburn symptoms for lansoprazole versus omeprazole.^{14, 15, 25} Although lansoprazole improve some symptoms faster at some time points, there was no strong or consistent pattern to suggest that lansoprazole provides faster symptom relief than omeprazole. Time to sustained resolution of heartburn (defined as 3 consecutive days without heartburn) was measured in one study and was similar (median 3 days for both drugs; p=0.285).¹⁴ In another study, daytime and nighttime heartburn were reported separately.²⁵ After one day of treatment, more lansoprazole-treated patients were free of both day heartburn (48.7% vs 37.6%; p<0.05) and night heartburn (62% vs 52%; p<0.05). The third comparison of these drugs used a visual analogue scale to measure heartburn symptoms and reported the time to relief only for daytime heartburn.¹⁵ After 3 days, there was a significant decrease in VAS score in lansoprazole-treated patients (−20.2 vs −15.3 (p=0.05); the difference was not significant after 7 days (scores not reported).

Rabeprazole vs omeprazole

One study reported similar mean time to complete heartburn relief for rabeprazole or omeprazole 20 mg daily; 7.2 and 8.4 days, respectively.³²

Esophagitis Healing

All of the PPIs were effective at healing esophagitis. Healing rates at 4 weeks ranged from 49% to 91%, and at 8 weeks ranged from 71 % to 99% (see Evidence Table 1). One small, fair quality study conducted at a single center in China had a lower 8-week healing rate than other studies (64% for esomeprazole 40 mg, 45.5% for omeprazole 20 mg).³¹

To determine an estimate of healing rates for each drug, we pooled data from head-to-head trials, using a random effects model to control for the effect of the study. Table 6 shows results of this analysis. (Note that for lansoprazole 15 mg, pantoprazole 20 mg, and rabeprazole 10 mg, these data are available from only one study). Healing rates were similar and confidence intervals overlapped, indicating no significant differences between PPIs.

Table 6

Estimated rates of esophagitis healing in head-to-head trials*.

We also calculated the percent risk difference for healing in head-to-head comparisons. Figures 1 and 2 show the differences in healing rates at 4 and/or 8 weeks for the 20 trials that provided the number healed/total patients.^{12, 14, 15, 17, 18, 20–23, 25–27, 30–32} Seven head-to-head trials are not represented in Figures 1 and 2; three studies (two of rabeprazole vs omeprazole, one of omeprazole vs both lansoprazole and rabeprazole)^{19, 28, 35} did not provide number healed/total, and four trials^{10, 11, 13, 16} reported only symptom relief, not esophagitis healing. The numbers presented in the figures for one trial of rabeprazole 20 mg compared to omeprazole 20 mg are calculated intention to treat, rather than those presented in the article which are not ITT.³²

Figure 1

Esophagitis healing rates at 4 weeks in head-to-head trials of PPIs (risk difference, 95% CI)

Figure 2

Esophagitis healing rates at 8 weeks in head-to-head trials of PPIs (risk difference, 95% CI)

Although some published studies presented results according to life-table analysis, crude rates only are included in figure 1. When a published study did not provide crude rates, we requested and received these data from the manufacturer. Results of life-table analyses cannot be directly compared to crude rates reported in other studies, and using life table analysis may overestimate results by excluding patients who are lost to followup or are withdrawn from the study.

Omeprazole 20mg, the first PPI to be marketed, was the comparator used most often in head-to-head trials. Table 7 summarizes the risk differences in healing rate in eight trials^{12, 15, 21, 22, 25–27, 31} comparing omeprazole 20 mg to another PPI.

Table 7

Risk differences of esophagitis healing rates in trials of omeprazole 20 mg vs another PPI*.

Risk differences at 4 and 8 weeks were non-significant in all comparisons, with the exception of esomeprazole 40 mg versus omeprazole 20 mg.

Two published trials comparing esomeprazole 40mg to omeprazole 20mg, found a higher healing rate in the esomeprazole group.^{6, 12} A third study³³ found no difference between groups at 4 and 8 weeks. A small study (N=48) found a higher healing rate for esomeprazole at 8 weeks (64% vs 45.5%), but the difference was not statistically significant.³¹ Four-week healing rates were not measured. This study may not have had sufficient power to detect a difference between treatment groups; no power calculation was reported. The pooled risk difference for three studies at 4 weeks was 8% and for 4 studies at 8 weeks was 6% (see Table 7). This translates to a number needed to treat to heal one additional patient at 4 weeks of 13, and a NNT at 8 weeks of 17.

Two studies compared esomeprazole 20mg to omeprazole 20 mg^{6, 7} and found no significant difference in healing rate at 4 weeks or 8 weeks.

Table 8 shows results of 3 studies that compared esomeprazole 40 mg to lansoprazole 30 mg. In a large, good quality trial in 5241 patients at multiple centers in the US,⁵ healing rates were higher in the esomeprazole group at 4 weeks (risk difference 4%; 95% CI 2%, 6%) and at 8 weeks (risk difference 3%, 95% CI 1%, 5%).

Table 8

Risk differences in head-to-head trials of esomeprazole 40 mg vs lansoprazole 30 mg.

A second, smaller, fair-quality trial of lansoprazole 30mg versus esomeprazole 40 mg¹⁸ in patients with mostly mild to moderate esophagitis found the two to have equivalent healing rates at 8 weeks. Results at 4 weeks are not reported.

The third study, rated good quality,²⁹ was conducted in patients with moderate to severe esophagitis (LA Grade C and D). At 4 weeks, the esomeprazole group had a higher healing rate, but at 8 weeks the difference was not significant.

Pooled estimates show a 5% higher healing rate at 4 weeks and 3% at 8 weeks for esomeprazole 40 mg. Using a random effects analysis, the difference at 8 weeks was not significant (95% CI 0%, 5%). In a fixed effects analysis, the difference is significant (risk difference 3%, 95% CI 2%, 5%). These estimates translate to a NNT to heal one additional patient at 4 weeks of 20; and at 8 weeks a NNT of 33.

Two trials compared esomeprazole 40 mg to pantoprazole 40 mg.^{20, 30} In one,³⁰ healing at 4 weeks was 6% higher at 4 weeks in the esomeprazole group (95% CI 3%, 9%). At 8 weeks, the difference was smaller but statistically significant (risk difference 3%; 95% CI 1%, 5%). We rated this study fair to poor quality. Data on baseline characteristics excludes 19 patients randomized but excluded from analysis due to intake of an unknown study drug or protocol violations. No data on excluded patients is presented. In addition, randomization and allocation concealment methods are not reported, and there were some differences in baseline esophagitis grade at baseline (grade B: 42.6% esomeprazole vs 45.1% pantoprazole; grade D: 4.5% esomeprazole, 5.8% pantoprazole).

In the other (fair-quality) comparison of esomeprazole 40 mg to pantoprazole 40 mg, healing rates are reported at "early" (4–6 weeks) and "late" (8–10 weeks) time points. Healing rates were equivalent at early and late time points. It was not possible to pool these studies because of the different manner of reporting results. Also, Gillessen included only patients with grade B (84%) and C (16%) esophagitis, whereas Labenz enrolled patients grade A through D.

Analysis of healing rates by baseline severity of esophagitis

Eighteen head-to-head trials reported information about esophagitis healing rates separately by baseline severity of esophagitis.^{5–7, 12–15, 18–23, 25, 27, 29, 30, 33} These results are shown in Evidence Table 1. Nine trials reported the number healed/total by baseline severity (Figures 4 and 5) Another reported a combined outcome of either healed or improvement by two grades.¹⁸

Figure 4

Healing at 4 weeks in patients with moderate to severe esophagitis (risk difference, 95% CI)

Figure 5

Healing at 8 weeks in patients with moderate to severe esophagitis (risk difference, 95% CI)

To estimate healing rates for each drug at 4 and 8 weeks for patients with moderate to severe esophagitis (i.e., grade C-D or 3–4; see Appendix F for scales used), we conducted a random effects meta-analysis of data from the 9 studies^{5–7, 12, 15, 25, 29, 30, 33} reporting the number healed/total by baseline severity (Table 9).

Table 9

Estimated healing rates in patients with moderate to severe esophagitis at baseline*.

Esomeprazole versus omeprazole

Three studies of esomeprazole 40 mg versus omeprazole 20 mg^{6, 12, 33} reported healing rates in patients with moderate to severe esophagitis at baseline (Figures 4 and 5). The pooled risk difference at 4 weeks was 16% (95% CI 11%, 22%) and at 8 weeks was 13% (95% CI 9%, 17%).

In two studies comparing esomeprazole 20 mg to omeprazole 20 mg^{6, 7} there was no difference in healing rate at 4 weeks (pooled risk difference 2%; 95% CI −5%, 10%) or 8 weeks (pooled risk difference 4%; 95% CI −3%, 10%). Estimates of healing rates in esomeprazole 20 mg are similar to omeprazole 20 mg (see Table 9). There are no comparisons of esomeprazole at any dose to omeprazole 40 mg.

Esomeprazole versus lansoprazole

Two studies of esomeprazole 40 mg versus lansoprazole 30 mg reported healing rates in patients with moderate to severe esophagitis at baseline.^{5, 29} The pooled risk difference at 4 weeks was 8% (95% CI 4%, 12%) and at 8 weeks was 9% (95% CI 5%, 12%). This corresponds to a NNT of 13 at 4 weeks and 11 at 8 weeks.

A third study, published by the maker of lansoprazole, reported only the combined outcome of healing or improvement of at least 2 grades in the subgroup of patients with moderate to severe esophagitis.¹⁸ In this study, there was a trend for a higher healing/improved rate in the lansoprazole group at 8 weeks (results at 4 weeks are not reported). The number of subjects in this subanalysis was comparatively small (N=109), and the difference was not statistically significant (10%, 95% CI −2%, 22%).

Esomeprazole versus pantoprazole

In one study, patients with moderate (Grade C) esophagitis at baseline taking pantoprazole 40 mg had a higher healing rate at "later" time points (8–10 weeks) than those taking esomeprazole 40 mg (67% vs 45%).²⁰ Among patients diagnosed with grade C at baseline, 100% of pantoprazole and 91% of esomeprazole improved to Grade A or B at the final visit (10 weeks). Rates at 4 weeks are not reported, and no patients with Grade D esophagitis were enrolled.

In the other comparison of esomeprazole 40 mg and pantoprazole 40 mg in patients with moderate to severe esophagitis, there was a 14% risk difference favoring esomeprazole after 4 weeks (95% CI 7%, 21%); 8-week data are not reported.

Lansoprazole versus omeprazole

Three studies comparing lansoprazole with omeprazole reported rates in patients with moderate to severe (Grades 3 and 4) esophagitis.^{14, 15, 25} Two of these compared lansoprazole 30 mg to omeprazole 20 mg.^{14, 15, 25} There was no difference in healing rate at 4 weeks (pooled risk difference 1%; 95% CI −13%, 16%) or 8 weeks (pooled risk difference 3%; 95% CI −4%, 10%). The third study compared lansoprazole 30 mg to omeprazole 40 mg and reported healing rates as percentages only.¹⁴ The number of patients with moderate to severe esophagitis in each group is not reported. There was no significant difference between groups at 4 or 8 weeks.

Non-erosive or empirically treated GERD

Direct Evidence

We identified 3 fair-quality head-to-head trials in patients with endoscopy negative reflux disease (ENRD, Table 10). They all compared esomeprazole to another PPI (omeprazole,³⁶ rabeprazole,³⁷ or pantoprazole³⁸). A fourth head-to-head trial of lansoprazole versus omeprazole included patients with both erosive and nonerosive esophagitis, but did not report results separately by these patient populations.³⁹

Table 10

Number of head-to-head trials, short-term treatment of non-erosive/ empirically treated GERD.

The three studies used different outcome measures, but all found esomeprazole to be similar in efficacy to the comparator (Evidence Table 3).

Three 4-week trials of omeprazole 20 mg versus esomeprazole 20 or 40 mg with identical designs were conducted simultaneously and are described in one publication.³⁶ There was no difference on the outcome of complete resolution of heartburn at 14 days (secondary outcome) or 28 days (primary outcome) between patients taking esomeprazole 20 mg or 40 mg, or omeprazole 20 mg. At 2 weeks, proportions of patients with complete resolution ranged from 34.6% to 44.3%, and at 4 weeks ranged from 56.7% to 70.3%. Results for adequate control of symptoms were similar, with no significant differences between drugs.

A head-to-head trial of pantoprazole 20 mg versus esomeprazole 20 mg measured time to first and sustained symptom relief.³⁸ This trial was designed to test the non-inferiority of pantoprazole compared with esomeprazole. The non-inferiority margin was set at −2 days for the primary outcome of time to first symptom relief (i.e., a lower boundary of the 95% confidence interval greater than 2 days would indicate non-inferiority). Symptom assessment was based on patient self-report using a validated questionnaire (ReQuest). The questionnaire includes the items on seven dimensions of GERD (general well-being, acid complaints, upper abdominal/stomach complaints, lower abdominal/digestive complaints, nausea, sleep disturbances, and other complaints). Results showed that pantoprazole was not inferior to esomeprazole for first and sustained symptom relief.

A 4-week trial of rabeprazole 10 mg compared to esomeprazole 20 mg was conducted in 134 patients in Singapore.³⁷ The primary outcome was time to achieve first 24-hour interval with no symptoms of heartburn or regurgitation. There was no difference between groups on this endpoint (8.5 days for rabeprazole vs 9.0 days for esomeprazole for heartburn; 6.0 days for rabeprazole vs 7.5 days for esomeprazole for regurgitation; p=NS;). There were also no significant differences between groups on secondary outcomes, including complete and satisfactory relief of heartburn symptoms at week 1 and week 4, and symptom severity score in the first 5 days.

Indirect Evidence

A good-quality Cochrane systematic review addressed the efficacy of PPIs, H2RAs, and prokinetics in adults with endoscopy negative or empirically treated reflux disease.⁴⁰ Literature searches were conducted through December 2003. This review was not designed to compare the efficacy of different PPIs, and included trials comparing PPIs to prokinetics, H2RAs, or placebo. The primary efficacy variable of the review was heartburn remission, defined as no more than one day with mild heartburn per week. PPIs were superior to placebo for heartburn remission and overall symptom improvement. PPIs were more effective than H2RAs for heartburn remission in empirically treated patients (pooled RR 0.69; 95% CI 0.61, 0.77), but not in patients with ENRD (pooled RR 0.74; 95% CI 0.53, 1.03). However, only 3 trials compared PPIs to H2RAs in ENRD.

Another systematic review evaluated the efficacy of PPIs for heartburn resolution in patients with ENRD.⁴¹ Searches for this review were conducted through 2002 and included the FDA web site. Seven placebo-controlled trials (three published and four unpublished) were included; 2 rabeprazole, 2 esomeprazole, and 3 omeprazole. In patients with ENRD, the risk difference versus placebo for complete resolution of heartburn at 4 weeks was 25% (95% CI 18% to 31%). This review does not provide evidence about comparative efficacy of different PPIs in patients with ENRD.

Table 11 shows the heartburn remission rates and rates of complete symptom relief calculated from data provided in the Cochrane review.⁴⁰ Similar proportions of patients experienced heartburn resolution or complete symptom relief across studies, with some exceptions in individual studies.

Table 11

Outcomes at 4 weeks, endoscopy negative/ empirically treated patients.

We identified one additional placebo-controlled⁴² and one active-controlled (vs. ranitidine) trial⁴³ published more recently and not included in this review (Evidence Table 3). In a fair-quality trial of empirical treatment of patients with symptoms of GERD, more patients taking pantoprazole 20 mg than ranitidine 300 mg were free of GERD symptoms (heartburn, acid eructation, and pain on swallowing) at 4 weeks (68.3% vs 43.3%).⁴³ In a fair to poor quality 8-week placebo-controlled trial of endoscopy-negative patients whose primary symptom was upper abdominal discomfort, patients taking lansoprazole 15 mg had fewer days with upper abdominal discomfort and reduced average daily pain severity.⁴² Patients whose predominant symptom was heartburn were not included. It is not clear what proportion of patients was analyzed; patients were excluded from analysis for a specific endpoint if there were no data available for that endpoint.

Prevention of Relapse

Patients with Erosive Esophagitis

Four randomized controlled trials compared one PPI to another for long-term (6 months or more) maintenance therapy for esophagitis relapse prevention in patients with endoscopically-proven GERD (Evidence Table 4).^{44–46, 47, Labenz, 2005 #4528} Two of these found no differences in endoscopic or symptomatic relapse rates for lansoprazole versus omeprazole after 48 weeks of treatment,⁴⁵ or rabeprazole versus omeprazole after 13 weeks, 26 weeks, one year, and five years.^{44, 47}

Two similar 6-month trials conducted by the same investigators compared esomeprazole 20 mg a day (an FDA approved dose for healing or maintenance of erosive esophagitis) to lansoprazole 15 mg a day (FDA approved dose for maintenance of healed erosive esophagitis) ⁴⁶ or pantoprazole 20 mg a day (lower than the FDA approved dose for maintenance of healed erosive esophagitis).⁴⁸ These studies randomized patients who had been healed after 4 to 8 weeks of treatment and compared relapse rates at 6 months. According to life-table analysis, a higher proportion of patients in the esomeprazole groups remained healed over 6 months: 83% vs 74% and 87% vs 74.9% esomeprazole vs lansoprazole and esomeprazole vs pantoprazole, respectively. The authors also present data by baseline severity. More patients in the esomeprazole groups remained healed across all grades of disease severity in both studies. The efficacy of lansoprazole and pantoprazole decreased with increasing severity of disease, while in only 1 study⁴⁸ the efficacy of esomeprazole in preventing relapse was lower in grade D patients compared to lower severity grades. No crude rates or numbers of patients remaining healed were presented. Crude rates provide a more conservative estimate of effectiveness due to the manner in which drop-outs are handled in life-table analyses. Because all patients enrolled in the study of esomeprazole and lansoprazole⁴⁶ had responded to esomeprazole for initial healing of esophagitis, the study may be biased towards esomeprazole. Both studies were funded by the manufacturer of esomeprazole and the publications were co-authored by representatives of the company.

A shorter-term trial of 36 patients with severe (Savary-Miller Grade 4) esophagitis compared omeprazole, lansoprazole, and pantoprazole for the prevention of relapse at 4 weeks.⁴⁹ Before randomization, all of the patients were treated with omeprazole. Six patients did not heal after 6 to 8 weeks of omeprazole; the rest (83%) were randomized to omeprazole, lansoprazole, or pantoprazole. After 4 weeks, patients taking omeprazole had a lower rate of endoscopic relapse (10%) than those randomized to either lansoprazole (80%) or pantoprazole (70%). The relapse rates in the lansoprazole and pantoprazole groups are very high compared with other studies and, as in the esomeprazole versus lansoprazole study discussed above, had a selection bias in that all subjects had responded well to one of the study drugs before enrollment in the maintenance phase.

A Cochrane systematic review estimated that the relative risk of relapse of healed esophagitis with a healing dose of a PPI compared with placebo was 0.26 (95% CI 0.19 to 0.36; NNT=1.7).⁵⁰ For maintenance of remission of symptoms, the relative risk was 0.34 (95% CI 0.25 to 0.46; NNT=2.0). For maintenance doses, the relative risk for remission of esophagitis was 0.46 (95% CI 0.38 to 0.57; NNT=2.4), and for remission of symptoms, the relative risk was 0.54 (95% CI 0.42 to 0.69; NNT=3.0). This review did not compare different PPIs.

Patients with non-erosive or empirically-treated GERD

We identified only one head-to-head trial of maintenance treatment in patients with non-erosive GERD.⁵¹ We also included two placebo-controlled trials of on-demand rabeprazole⁵² and esomeprazole,⁵³ and a placebo-controlled trial of continuous omeprazole.⁵⁴ Details of these trials are shown in Evidence Table 5. Three other trials included patients with both endoscopy negative GERD and erosive esophagitis, but did not report results separately by group.^{34, 55, 56}

A head-to-head trial compared esomeprazole 20 mg on-demand to continuous lansoprazole 15 mg for 6 months in patients with ENRD who had experienced complete relief of heartburn with esomeprazole 20 mg during an acute treatment phase (2 to 4 weeks).⁵¹ Patients were not blinded to treatment, and the primary outcome measure was time to discontinuation from the maintenance phase due to unwillingness to continue. Patients also recorded heartburn and other symptoms on diary cards and were asked about their satisfaction with treatment during scheduled clinic visits. By 6 months, significantly more patients receiving continuous lansoprazole 15 mg were unwilling to continue compared to those receiving esomeprazole 20 mg on-demand (13% vs 6%, p=0.001). More patients in the lansoprazole group said they discontinued because of adverse events (7.4% vs 2.3%; p=0.0028), but discontinuations because of heartburn were not significantly different between treatment groups (2.9% vs 4.8% for esomeprazole and lansoprazole, respectively; P not reported, but NS). At 1 month, more esomeprazole patients were satisfied with their treatment, but at 3 and 6 months, there was no difference between treatment groups on this measure. During the maintenance phase, the mean frequency of heartburn symptoms was higher in the esomeprazole on-demand group compared with the continuous lansoprazole group.

Two 6-month placebo-controlled studies reported efficacy of on-demand therapy with rabeprazole 10 mg⁵² or esomeprazole 20 mg⁵³ in patients with ENRD. In both studies, only patients who experienced complete symptom relief during an acute treatment phase were enrolled in the maintenance phase. In the study of rabeprazole 10 mg, rates of discontinuation due to inadequate heartburn control were 20% for placebo versus 6% for rabeprazole (p<0.00001). Although mean duration of heartburn-free periods was similar between groups, the time required for symptom resolution during a heartburn episode was significantly shorter with rabeprazole than placebo. In the study of esomeprazole 20 mg, 14% of patients taking esomeprazole versus 51% of those taking placebo discontinued, mainly due to inadequate control of heartburn (p<0.0001).⁵³

In a placebo-controlled trial of continuous omeprazole 10 mg, 27% of patients taking omeprazole versus 52% of those taking placebo discontinued due to inadequate control of heartburn over 6 months.⁵⁴

Systematic reviews of head-to-head trials in patients with erosive esophagitis

Four recent systematic reviews comparing PPIs for esophagitis healing and symptom relief have been published.^57–60 Three of the four included studies of esomeprazole, and all concluded that esomeprazole was superior to other PPIs for GERD, based on the same studies included in this report.^58–60 One of these concludes that better healing rates in patients taking esomeprazole 40 mg compared with those taking omeprazole 20 mg or lansoprazole 30 mg is attributable to increased efficacy of esomeprazole in patients with more severe grades of esophagitis.⁵⁸ The other was designed to compare the efficacy of esomeprazole versus lansoprazole, and concluded that esomeprazole provided an additional benefit of 5% at 4 weeks and 4% at 8 weeks compared with lansoprazole 30 mg.⁶⁰ Both of these were funded by the manufacturer of esomeprazole.

A third systematic review,⁵⁹ in which the funding source is not reported, concluded that esomeprazole 40 mg was superior to omeprazole 20 mg for GERD healing after 4 weeks (RR 1.18, 95% CI 1.14–1.23), but that this result was due to the non-equivalent, higher dose of esomeprazole used. There were no differences among the other PPIs.

A systematic review conducted in 2001⁵⁷ found that lansoprazole, rabeprazole, and pantoprazole had similar efficacy to omeprazole for healing. No study of esomeprazole had been done at the time.

Summary of the Evidence

Duodenal ulcer:

• The data regarding comparative effectiveness of various PPIs for treating duodenal ulcer are good, with nine head-to-head trials.
• Omeprazole 20mg daily is typically the comparator drug.
• The evidence is good for omeprazole and lansoprazole having similar effectiveness in both endoscopic healing and symptom relief. The pooled risk difference for five trials of lansoprazole 30mg versus omeprazole 20mg once daily is −0.2 (95% CI, −3.0 to +2.6).
• The evidence for pantoprazole, rabeprazole and esomeprazole is less strong, because there are only single studies for each drug compared to another PPI (all compared to omeprazole).
• No study found significant differences in healing rate. Data from studies comparing PPIs to H2-RAs also indicate that there are no significant differences between the four PPIs studied (there are no studies of esomeprazole).
• Symptom relief is an important measure in ulcer diseases, and does not always correspond to endoscopic healing. Method for assessment of symptom relief was not consistent across the studies, and reporting of findings was often limited to early time periods and just a few outcome measures (of many measured). Few studies found a difference in any of the many measures of symptom relief, and the lack of reported data at later time-points may indicate that symptom relief was equivalent.

Gastric ulcer:

• Comparative data about PPIs for the treatment of gastric ulcer is very limited, with 2 studies of rabeprazole versus omeprazole. No significant differences in healing rates were found.
• Data from studies of omeprazole, lansoprazole and pantoprazole compared to H2-RAs indicate no significant difference in the rate of healing at 4 weeks.
• Symptom relief was better in 3 of 12 measures for rabeprazole compared to omeprazole at 3 weeks or two measures and 6 weeks for a third measure (the measures significantly different at 3 weeks were not different at 6 weeks). Symptom relief was difficult to compare for the other drugs, with no head-to-head studies. No important difference was clear from the PPI versus H2-RA studies.

NSAID-induced ulcer:

• There are no head-to-head trials.
• Only 4 trials compared a PPI to another drug, two with omeprazole one each with esomeprazole and lansoprazole. No important differences between PPIs could be discerned from these studies, with the confidence intervals for healing rates overlapping. However, the treatment success rates for all treatments varied widely among the trials, so confidence in this finding is low.

Prevention of NSAID-induced ulcer:

• There are no head-to-head trials.
• A good quality systematic review and seven subsequently published trials compared PPIs to placebo or other drugs. Only one trial included outcome measures for serious ulcer complications, and for some of the endoscopic ulcer findings, patients were asymptomatic.
• Based on development of new ulcers or serious erosions and on symptoms, there did not appear to be differences in the PPIs studied (omeprazole, lansoprazole and pantoprazole). However, because of the differences in patient populations, comparison groups, and outcome measure definitions, confidence in this finding is low.

Eradication of H. pylori:

• The evidence regarding comparative effectiveness of various PPIs is fair, with five systematic reviews, and 25 recent head-to-head trials. The significant heterogeneity among studies based on design, participants, and method of measuring outcomes lessen the strength of the evidence.
• These studies generally did not find a difference in eradication rate between the PPIs, with the exception of lower dose pantoprazole when compared to high dose pantoprazole or high dose omeprazole, pantoprazole compared to esomeprazole in one study, and rabeprazole when compared to lansoprazole in one study.
• Symptom resolution was not assessed in these studies.
• In children, evidence is extremely limited with only 2 trials of lansoprazole versus placebo. Neither trial found the addition of lansoprazole to result in higher eradication rates than antibiotic therapy alone.

Detailed Assessment

Key Question 2a. In head-to-head comparisons, what is the comparative efficacy of different PPIs in reducing symptoms and improving endoscopic healing in patients with duodenal ulcer?

Nine randomized controlled trials compared one PPI to another.^{35, 89–96} The details of these studies are summarized in Evidence Table 7. Six of these trials compared lansoprazole 30mg to omeprazole 20mg.^{89–93, 96} One study each compared pantoprazole 40mg and rabeprazole 20mg to omeprazole 20mg^{35, 94} and one study comparing esomeprazole 40mg to omeprazole 40mg.⁹⁵ All of these dose comparisons are fair based on equipotency.

The studies were fair quality. These studies were generally similar with respect to design, demographics and other population characteristics, with the following exceptions. One study was unusual in that as a part of a H. pylori eradication regimen, patients with active duodenal ulcer were given esomeprazole plus antibiotics for only 1 week, while omeprazole patients received antibiotics plus omeprazole for 1 week, then continued omeprazole for another 3 weeks.⁹⁵

As shown in Figure 3, there was no difference between omeprazole 20mg, lansoprazole 30mg, and rabeprazole 20mg in the percentage of patients healed by 4 weeks. Results from a large multicenter trial of esomeprazole 20mg twice daily versus omeprazole 20mg twice daily also showed no difference in healing rates.⁹⁵ The pooled risk difference for lansoprazole 30mg versus omeprazole 20mg once a day was −0.2 (95% CI, −3.0 to +2.6). The risk differences found between esomeprazole 40mg, pantoprazole 40mg and rabeprazole 20mg and omeprazole were approximately −0.97%, 6% and 5%, respectively, however these are based on single studies and were not statistically significant. The results for healing at 2 weeks were similar.

Symptoms (pain, nausea, vomiting, antacid use, or overall well-being) were assessed by investigators at visits and through patient diaries in seven studies. Only one found a significant difference between PPIs.³⁵ This study found that daytime pain was `improved' in 92% on rabeprazole and 83% on omeprazole at 4 weeks (p=0.038), however no difference was found in nighttime pain or in the number of patients who were pain-free. Antacid use, GI symptoms, and overall well-being were not different in any of the studies.

Only one head-to-head study addressed maintenance, comparing lansoprazole 15mg, lansoprazole 30mg and omeprazole 20mg for up to 12 months (see Evidence Table 8).⁹² At 6 months post-healing, recurrence rates were 4.5%, 0%, and 6.3%, respectively. At 12 months the recurrence rates were 3.3%, 0%, and 3.5%, respectively. These differences were not statistically significant.

Three other studies listed in Evidence Table 8 compared lansoprazole to placebo^{97, 98} or ranitidine.⁹⁹ Relapse rates at 12 months in the lansoprazole 15mg groups ranged from 23 to 30%, in the single lansoprazole 30mg group the rate was 15%, compared to placebo rates of 39 to 100%. One study reported relapse rates with no maintenance treatment following healing with omeprazole, ranitidine or placebo. Relapse rates were not significantly different between the groups.

Key Question 2b. In comparisons of PPIs and H2-RAs, what is the comparative efficacy of different PPIs in reducing symptoms and improving endoscopic healing in patients with duodenal ulcer?

Twenty-five randomized controlled trials compared a PPI with an H2-RA. Of these, 22 papers were reviewed.^100–121 Since these studies can only be used to make indirect comparisons of the effectiveness of the various PPIs, a limited analysis is presented. Individual study quality assessments for these studies will not be presented. If an obvious difference in healing rate were seen in an individual study or studies, investigation of study quality would have been undertaken.

The most common H2-RA used as a comparator was ranitidine 300mg per day, with ten studies comparing omeprazole 20mg, four studies comparing pantoprazole 40mg, two studies comparing lansoprazole (doses varying from 15 to 60mg per day), and one study comparing rabeprazole 20mg. Two compared omeprazole 20mg to cimetidine (doses varying from 800mg to 1200mg per day), two compared omeprazole 20mg with famotidine 40mg, and 1 compared omeprazole with nizatidine 300mg. There are no studies comparing esomeprazole to an H2-RA.

Figure 4 shows the rates of duodenal ulcer healing at 4 weeks in 21 studies of a PPI versus an H2-RA PPIs were more effective at healing than H2-RAs, but there were no significant differences in healing rates among the PPIs. Duodenal ulcer healing rate at 4 weeks with omeprazole and lansoprazole was dependent on H2-RAs healing. That is, as the healing rate in the H2-RA group increased, PPI healing rate increased. One comparison showed pantoprazole to have a significantly higher healing rate than rabeprazole (risk difference 11.3%), but this comparison is based on only one study, and the confidence interval is large (95% CI, 2.4%–23.2%).

Another study¹²² examined the added benefit of continuing omeprazole 20 mg for 3 additional weeks after 1 week of eradication therapy with omeprazole 20mg combined with amoxicillin 1000 mg and clarithromycin 500 mg. At 4 weeks, there was no difference in healing rates in patients assigned to omeprazole (89%) versus placebo (87%). An additional four trials were found in updating the original review^{121, 123–125} These studies were consistent with the studies reported above and are not added to figure 4. One of these studies reported symptom relief only.¹²¹

Key Question 2c. In head-to-head comparisons, what is the comparative efficacy different PPIs in reducing symptoms and improving endoscopic healing in patients with gastric ulcer?

Three studies compared PPIs directly in treating gastric ulcer.^126–128 A fair quality study of 227 patients compared rabeprazole 20mg to omeprazole 20mg (Evidence Table 9).¹²⁶ Healing was assessed at 3 and 6 weeks, while most other studies of gastric ulcer healing use 4 and 8 weeks. The percent risk difference in the rate of healing at 3 weeks is −3% (95% CI, −16, 9.7), and reported as the same in both groups at 6 weeks. Symptoms were assessed by investigators at visits and through patient diaries. Twelve different comparisons of symptom resolution or improvement were made. No significant differences were found in the reporting of pain resolution or improvement (frequency, severity, night or daytime) at 3 or 6 weeks for nine of these comparisons. Rabeprazole was statistically superior in three comparisons: improvement of severity of pain at 3 weeks and improvement in the frequency of daytime pain and resolution of nighttime pain at 6 weeks. No difference in changes in overall well-being or reduction in antacid use was found.

A smaller fair quality study (n = 80) compared rabeprazole 10 mg a day and omeprazole 20 mg a day, and evaluated the impact of CYP2C19 genotype on healing rates.¹²⁸ The overall healing rate comparison at 8 weeks (risk difference 1.94%, 95% CI −1.34% to 1.71%) did not show a difference between the drugs.

A poor quality trial compared lansoprazole 30 mg/day to omeprazole 20 mg/day. This study did not conduct and intent to treat analysis and more patients were excluded from the omeprazole (15%) analysis than the lansoprazole group (7%). Although the authors state there were no differences between groups at baseline, 4% of patients in the omeprazole group were smokers, compared to 1% in the lansoprazole group. The results of this study found lansoprazole superior in cumulative healing rate at 8 weeks (93% vs 82%, p=0.04); the difference at 4 weeks was not statistically significant. It is not clear from the publication which patients were included in this analysis, and our statistical analyses based on differing assumptions did not result in statistically significant differences between the groups at either time point. Differences in symptom relief were not found to be statistically significant.

Key Question 2d. In comparisons of PPIs and H2-RAs, what is the comparative efficacy of different PPIs in reducing symptoms and improving endoscopic healing in patients with gastric ulcer?

Fifteen studies compared a PPI to an H2-RA for treatment of gastric ulcer (Evidence Table 9).^{93, 100, 129–141} There were two studies of maintenance therapy and one followup study of relapse rates in patients healed in one of the above studies.^{98, 142, 143} One of the maintenance studies included patients with either gastric or duodenal ulcer, all of which were resistant to H2-RA therapy.¹⁴³ One study evaluated esomeprazole versus ranitidine in healing ulcers in patient who continued to take a NSAID.¹²⁹ This study is examined below, in key question 2f. No study compared rabeprazole to a H2-RA. Five trials compared omeprazole to ranitidine; three compared lansoprazole to ranitidine; one compared pantoprazole to ranitidine; two, lansoprazole to famotidine; three, omeprazole to cimetidine, and one, lansoprazole to cimetidine.

The total followup times varied, but healing rates at 4 weeks were available from all studies. Differences in the percentages of patients healed with different PPIs at 4 weeks are plotted in Figure 5 The pooled risk differences range from 1.09 to 62.5%, with the smallest studies showing larger effects. The confidence intervals for PPIs compared to H2-RAs all overlap.

Symptoms were assessed by investigators at visits and through patient diaries in 13 studies. One did not report symptoms.¹³² Pain was the most commonly assessed symptom. The scales used were not consistent across the studies (0 to 3 in some, 0 to 4 in others), or were not described. Most found the PPI relieved symptoms somewhat faster, with no difference later on. However, only three studies found statistically significant differences, and then only in some of the many measures assessed.

One study¹⁴⁴ reported maintenance therapy of lansoprazole 15 or 30mg compared to placebo. Lansoprazole was effective for preventing endoscopic recurrence and eliminating symptoms and reducing antacid use. Omeprazole 20 mg every day was more effective than ranitidine in preventing relapse in patients with refractory ulcer (not healed after 8 weeks of H2-RA treatment) in one 6-month open study.¹⁴³ Only 12 patients of 102 enrolled were assigned to ranitidine in this study, and patients with both gastric and duodenal ulcer were included. A 6-month followup study without treatment¹⁴² of patients who had healed after 6 weeks of treatment with omeprazole or cimetidine¹³¹ found no significant difference in relapse rates. All of these studies had high or differential dropout rates.

Key Question 2e. In head-to-head comparisons, what is the comparative efficacy of different PPIs in reducing symptoms and improving endoscopic healing in patients with NSAID-induced ulcer?

No study compared one PPI to another.

Key Question 2f. In comparisons of PPIs and misoprostol or H2-RAs, what is the comparative efficacy of different PPIs in reducing symptoms and improving endoscopic healing in patients with NSAID-induced ulcer?

Four studies assessed PPIs (omeprazole, esomeprazole, and lansoprazole) compared to another drug in healing ulcers induced by NSAIDs.^{129, 145–147} The details of these studies are summarized in Evidence Table 10. A good quality systematic review of prevention and treatment of NSAID induced ulcers was also found.¹⁴⁸

Comparisons of ranitidine 150 mg twice daily to omeprazole 20 and 40 mg daily, lansoprazole 15 and 30 mg daily and esomeprazole 20 and 40 mg once daily resulted in higher rates of healed ulcer at 8 weeks for the PPI.^{129, 145, 147} The difference in percent healed ranged from 14 to 22% favoring the PPI, but in all comparisons the difference was statistically significant. While there is no direct comparison of the PPIs, all confidence intervals overlap. A single study found that 20 mg of omeprazole was superior to misoprostol in healing rates at 8 weeks, but that 40 mg was not superior.¹⁴⁶

One study^{146, 149} assessed quality of life using the Gastrointestinal Symptom Rating Scale and the Nottingham Health Profile. Based on the Gastrointestinal Symptom Rating Scale, omeprazole was better than misoprostol in the changes in scores for the total scale, as well as scores for reflux and diarrhea. Although the improvement in score was greater with 20mg omeprazole than 40mg, these were not statistically significant. Only the sleep score of the Nottingham Health Profile was reported, which also showed omeprazole 20mg to be superior to misoprostol, but the change in score for omeprazole 40mg was not reported.

Key Question 2g. In head-to-head comparisons, what is the comparative efficacy of different PPIs in preventing NSAID-induced ulcer?

There are no head-to-head comparison studies.

Key Question 2h. In comparisons of PPIs, other drugs, or placebo what is the comparative efficacy of different PPIs in preventing NSAID-induced ulcer?

One recent, good quality systematic review addressed this question.¹⁵⁰ The search for literature covered 1966 to 2000 (MEDLINE search from 1966 to January 2000, Current Contents for 6 months prior to January 2000, EMBASE to February 1999, and a search of the Cochrane Controlled Trials Register from 1973 to 1999). This review found five randomized trials, which assessed omeprazole 20 to 40mg in prevention of NSAID-induced gastroduodenal toxicity. None of the studies were designed to evaluate the effectiveness of PPIs in preventing serious ulcer complications (hemorrhage, perforation or death). The review showed that omeprazole is superior to the H2-RAs but provided no data on any other PPI.

Five trials published more recently^151–155 are presented in Evidence Table 11, along with two of the treatment studies that included a prevention phase.^{146, 147} None of these studies was a head-to-head comparison and there were important differences in treatment regimens and followup, making comparisons across studies impossible. All enrolled patients who were regular users of NSAIDs. One study¹⁵¹ included only patients who were H. pylori negative and randomized to placebo, misoprostol 800mcg, lansoprazole 15mg or 30mg with followup at 1,2 and 3 months, another¹⁵² randomized patients to pantoprazole 40mg or placebo for 3 months. The third study¹⁵³ included patients who were H.pylori positive and had ulcer complications after using low-dose aspirin continuously for more than one month. After ulcers were healed and H. pylori eradicated, patients were randomized to lansoprazole 30 mg or placebo, in addition to 100 mg of aspirin daily. In the last study,¹⁵⁴ H.pylori positive patients with no past or current ulcer were assigned to one of 4 treatment groups: omeprazole 20 mg plus clarithromycin 500 mg and amoxicillin 1 gram for one week, followed by placebo or omeprazole 20 mg daily for 4 weeks; omeprazole 20 mg once daily for five weeks; or placebo for 5 weeks.

In the study of H. pylori negative patients,¹⁵¹ lansoprazole was inferior to misoprostol in preventing gastric ulcers. At 3 months, the gastric ulcer rate (failure rate) was 7% for misoprostol, 20% for lansoprazole 15mg, and 18% for lansoprazole 30mg, with no significant difference between lansoprazole doses. However, when adverse effects were included as failures, the failure rate for all 3 treatment groups was 31%. A post-hoc subgroup analysis of patients taking NSAIDs and low dose aspirin found no significant differences were found among the groups of drug treatments at 12 weeks.¹⁵⁶

In the study of pantoprazole versus placebo,¹⁵² a life-table analysis is presented, rather than simple proportions of patients without ulcer, making comparison to other PPI versus placebo studies unclear. At 4 weeks, the risk difference is 17% fewer ulcers in the pantoprazole group, and 27% at 12 weeks. These numbers include those who dropped out due to adverse effects as treatment failures.

In the study of H.pylori positive patients with ulcer complications,¹⁵³ the primary endpoint was prevention of ulcer complications and the secondary endpoint was recurrence. The rate of recurrence of ulcer complications at a median followup of 12 months was 1.6% in the lansoprazole group, compared with 14.8% in the placebo group. Two patients in the placebo group were also taking NSAIDS.

In patients with H.pylori but no history of ulcer, all 3 active treatment regimens were better than placebo in reducing the occurrence of ulcer and dyspeptic symptoms requiring therapy, and there were no significant differences between the treatment groups.

Symptom assessment and reporting varied among these studies. The pantoprazole versus placebo study did not describe methods or scales used to assess symptoms, but reported "GI symptoms."¹⁵² GI symptoms were not the same at baseline in the two groups; 43% in the pantoprazole versus 18% in placebo group complained of GI symptoms. At 4 and 12 weeks the pantoprazole group improved (17% and 20%, respectively), while the placebo group remained stable (20% and 19%, respectively). In the lansoprazole versus misoprostol study, symptoms (day and nighttime abdominal pain and antacid use) were assessed by patient diary and were found to be significantly better in the lansoprazole groups versus misoprostol, but comparisons between the two lansoprazole doses were not made.¹⁵¹

Key Question 2i. In head-to-head comparisons, what is the comparative efficacy of different PPIs in improving eradication rates in patients with Helicobacter pylori?

A good-quality meta-analysis reviewed 14 head-to-head trials of PPIs combined with antibiotics in triple-therapy regimens for h. pylori eradication.¹⁵⁷ Using omeprazole as the reference for comparison, no difference was found in eradication rates among any of the PPIs. In addition, a fair quality systematic review addressed this question.¹⁵⁸ The search for literature covered 1986 to1998 (MEDLINE search from 1986 to 1997, and hand searches from 1986 to January 1998). This meta-analysis included 666 studies overall. Although the number of studies evaluating a PPI is unclear, there were nine different regimens that included a PPI. The PPIs included in these studies were omeprazole, lansoprazole, and pantoprazole. Using a meta-regression analysis, no difference in cure rate was found between the three PPIs in any of the antibiotic combinations studied. Another recent fair quality systematic review focused on lansoprazole in eradication of H. pylori.¹⁵⁹ This review found no difference between lansoprazole and omeprazole in eradication rate.

Since these reviews, 25 studies were published that directly compared one PPI to another in combination with the same antibiotic(s).^{95, 96, 160–179} They made the following comparisons:

• Rabeprazole 20mg versus omeprazole 40mg, plus amoxicillin (one study)¹⁷¹
• Lansoprazole 60mg versus omeprazole 40mg, plus amoxicillin and metronidazole (one study)¹⁶⁴
• Omeprazole 40mg versus pantoprazole 40mg, plus clarithromycin and metronidazole (one study)¹⁶⁰
• Omeprazole 20mg versus lansoprazole 30mg, plus clarithromycin and tinidazole (one study) ⁹⁶
• Various doses of lansoprazole, rabeprazole, pantoprazole and esomeprazole versus omeprazole, plus clarithromycin and amoxicillin (ten studies)^{76, 95, 161, 163, 167–169, 174, 177, 180–182}
• Omeprazole 20 mg, lansoprazole 30 mg, or rabeprazole 10mg (all twice daily) each combined with amoxicillin and clarithromycin (one study),¹⁶⁵
• Rabeprazole 10 mg or 20mg or lansoprazole 30mg twice daily, each combined with amoxicillin and clarithromycin (three studies),^{136, 166, 170, 172}
• Lansoprazole 30 mg or omeprazole 20 mg twice daily combined with amoxicillin alone, versus lansoprazole 30 mg twice daily combined with amoxicillin and clarithromycin (one study).¹⁷³
• Lansoprazole 30mg once a day vs omeprazole 20 mg twice a day combined with amoxicillin and clarithromycin. (1 study).¹⁷⁶
• Lansoprazole 30 mg or omeprazole 20 mg (both twice daily) plus amoxicillin and clarithromycin (1 study).¹⁶²
• Lansoprazole 30 mg or omeprazole 20 mg (both twice daily) plus amoxicillin alone (1 study).¹⁸³
• Esomeprazole 40mg versus omeprazole 40mg, plus clarithromycin and metronidazole (one study)¹⁸⁴
• Esomeprazole 40 mg daily, esomeprazole 40 mg twice daily, or omeprazole 20 mg twice daily plus amoxacillin and clarithromycin (1 study).¹⁷⁸
• Esomeprazole 40 mg or pantoprazole 40 mg (both once daily) plus amoxicillin and clarithromycin (w study).¹⁷⁹

These studies were fair quality, with the exception of 5 poor quality studies that were not blinded.^{171, 176, 178, 182, 184} This is a heterogeneous group of studies. Some of the PPI comparisons did not use what would be considered equivalent doses (e.g., rabeprazole 20mg versus omeprazole 40mg or omeprazole 40mg versus pantoprazole 40mg) and one used a dose of omeprazole that is not standard in the US (60mg).¹⁶⁷ In addition, the doses of clarithromycin, amoxicillin and metronidazole also varied. Some of the studies were assessing short durations of treatment; while others were evaluating the use of lower doses of PPIs in Asian patients (see Key Question 3). The methods of assessing H. pylori eradication also varied among the studies, as did other treatments during the study period. Hence, direct comparison across all studies is not possible.

Twelve studies included patients with documented ulcer.^{95, 96, 161, 162, 164–166, 169, 171, 172, 175–177, 182} Nine studies included patients with ulcers or non-ulcer dyspepsia^{76, 160, 163, 168, 170, 174, 181, 184} The proportion of non-ulcer patients ranged from 12%¹⁷⁰ to 71%.¹⁷⁴ One study conducted in a low-income population in Colombia included patients with "gastritis" and did not check for ulcer,¹⁶⁷ and two included both patients with previous or present recurrent ulcer.¹⁷³

As would be expected based on these differences, eradication rates varied in these studies, from a low of 42% (lansoprazole 30mg with only amoxicillin)¹⁷⁵ to a high of 100% (pantoprazole 40mg).¹⁶⁰ One study found a significantly lower eradication rate for pantoprazole (40mg) than for higher relative doses of omeprazole (40mg) or high-dose pantoprazole (80mg). Another study found a lower eradication rate for pantoprazole 40 mg compared to esomeprazole 40 mg daily¹⁷⁹. Rabeprazole (20 mg or 40 mg) had lower rates of eradication than lansoprazole 30 mg in another comparison.¹⁷² No other study found a significant difference regardless of dose or specific PPI. A study evaluating the effect of age found that older patients (> 50 years) had higher eradication rates than younger patients.¹⁷⁵

Key Question 2j. In comparisons of PPIs and H2-RAs, what is the comparative efficacy of different PPIs in improving eradication rates in patients with Helicobacter pylori?

Summary of the Evidence

• The comparative evidence on long-term adverse effects is limited. There are no long-term head-to-head comparative studies (clinical or observational) specifically designed to monitor adverse effects.
• Two long-term (48 weeks to 5 years) maintenance studies found no difference between omeprazole and lansoprazole in adverse events or withdrawals due to adverse events, and a 6-month study of esomeprazole 20 mg versus lansoprazole 15 mg found no differences in adverse event rates.
• In long-term followup studies of individual drugs, no important differences in long-term findings were apparent, but comparisons across these studies are not clear.
• Short-term head-to-head comparative studies indicate that the incidence of all and serious adverse events, and the drop out rate due to adverse events for all the PPIs is low. No consistent differences between the PPIs were seen in these trials.
• All PPIs share drug interactions based on elevated gastric pH altering absorption of a small number of drugs. Omeprazole is known to have drug interactions with a small number of drugs metabolized by the CYP2C19 and CYP3A4 enzyme systems. The action required is monitoring to see if dose adjustment of the other drug(s) is necessary. Lansoprazole may possibly interact with theophylline. Pantoprazole, rabeprazole, and esomeprazole have no documented drug interactions deemed clinically significant.

Detailed Assessment

There are no head-to-head long-term comparison studies designed to assess adverse events between PPIs. In three long-term (6 months or longer) maintenance studies of patients with GERD,^{45, 47, 190} there was no difference in the number of adverse events reported or number of withdrawals due to adverse events in the different PPI treatment groups. In one study of GERD patients,⁴⁵ 9 of 248 (3.6%) patients withdrew for adverse events over 48 weeks of treatment, 4% in the lansoprazole group and 3.3% in the omeprazole group. In another study, comparing rabeprazole 10 or 20mg to omeprazole 20mg 13 of 243 (5.3%) patients withdrew because of adverse events at 52 weeks,⁴⁴ and 26 of 243 (11%) withdrew at 5 years;⁴⁷ the numbers in each group did not differ significantly. In the third long-term maintenance study,¹⁹⁰ 29 of 617 (4.7%) patients in the esomeprazole 20 mg group and 32/614 (5.2%) of those in the lansoprazole 15 mg group withdrew due to adverse effects. There are no head-to-head maintenance studies of ulcer, but three 12-month studies of duodenal ulcer maintenance compared a PPI to placebo or other anti-ulcer medications. In two of the studies, the withdrawal rates for placebo were higher than any of the drug arms. In one study, the withdrawal rates due to adverse events were high, 17% for lansoprazole 15mg, 5.3% for lansoprazole 30mg and 21.5% for placebo over a 12-month period.⁹⁸

Several reports of long-term (ranging from 1 year up to 11 years) followup of individual PPIs (omeprazole, lansoprazole, pantoprazole, and rabeprazole) have been published.^191–206 Potential adverse events studied include hypergastrinemia related enterochromaffin-like cell (ECL) hyperplasia and ECL carcinoids, atrophic gastritis and intestinal metaplasia, overgrowth of gastric bacteria and N-nitrosamine formation, enteric infections, potential malabsorption syndromes, and diarrhea. Of these, the risk of enteric infections may be increased with sustained acid suppression. This is a rare event, however. The other concerns have not been proven in these long-term, non-comparative studies. While ECL hyperplasia occurs, no increased risk of ECL carcinoids has been found. Likewise, atrophic gastritis is increased with long term PPI therapy, but progression to intestinal metaplasia and gastric cancer has not been shown. Gastric bacterial overgrowth does occur, but a related higher rate of gastric adenocarcinoma has not been found. Long-term studies assessing the risk of esophageal cancer were not found.

A nested case-control study of 10,008 lansoprazole users followed for 4 years found a trend for diarrhea to be dose related, reported in 5%, 3.7%, and 2.5% of patients using 60 mg or more, 30 mg, and 15 mg or less, respectively (p=0.08). In 42.1% of patients reporting diarrhea the lansoprazole dosage was reduced or discontinued due to this event. Cases had a higher current use of oral antibiotics than controls with no diarrhea (adjusted OR 2.7, 95% CI 1.0–6.9). There are no long-term studies of esomeprazole or rabeprazole.

Reports of adverse effects in head-to-head comparisons of PPIs for short-term treatment of GERD and ulcer are shown in Evidence Table 12. The proportion of patients withdrawing due to adverse events in these studies was very low, with most studies reporting 1% to 3%. No study found significant differences among treatment groups in the rate of withdrawals for adverse effects. Reports of serious adverse events were low, and generally balanced among the drugs. Many of these incidences could be associated with pre-existing diseases.

Serum gastrin levels were monitored in several studies, and found to be significantly elevated compared to baseline although the magnitude of increase was small and generally not considered clinically significant. A dose-related difference was found in some studies, but no differences between drugs. Likewise, when studied, the effect of the individual PPIs on H. pylori-related gastritis was similar, worsening gastritis in the corpus, and improving gastritis in the antrum.²⁰⁷

Also in Evidence Table 12 is a head-to-head study designed to determine patient preferences about switching from one PPI to another.²⁰⁸ The study included patients who had been taking a PPI for any indication for at least 56 days before the start of the study. All patients took omeprazole 20 mg and rabeprazole 20 mg daily for 4 weeks in a crossover design, with the order of medication randomized. A double-dummy presentation was used to blind patients to treatment assignment. At the end of each 4-week treatment phase patients were asked to name any unwanted or welcome side effects from the medication. The two PPIs maintained similar relief of symptoms, and the tolerability was similar.

Drug Interactions

There are no head-to-head comparative studies of drug interactions with PPIs in patients with acid-related diseases. Drug interaction studies in healthy adults have been done with individual PPIs, and are summarized in Table 12, below. All of the PPIs reduce the absorption of drugs that require an acidic gastric pH for maximal absorption, such as ketoconazole. With all of the PPIs, the dose of these drugs may need to be increased, or the drug combination avoided (e.g., delaviridine and PPIs). All of the PPIs are metabolized by the CYP2C19 and CYP3A4 enzyme systems, and have some potential for interacting with other drugs that are also metabolized through this pathway. As can be seen in the table, omeprazole interacts with several drugs, but only four require any action (carbamazepine, phenytoin, diazepam and trovafloxacin). The recommended action is to monitor the patient for signs of adverse effects due to increased levels of these drugs. The newer PPIs have fewer studies of drug interactions, but in the studies that have been done, no clinically significant drug interactions have been found. The one possible exception to this is the decreased clearance of theophylline with lansoprazole. Since these studies have been done in healthy people, the external validity of the judgment of no clinical significance is unknown.

Table 12

Clinically significant drug interactions.

Summary of the Evidence

• Head-to-head comparison studies did not adequately describe or analyze subgroups for differences in effectiveness, although two assessed differences in adverse effects based on age, gender and race with no differences found.
• There are studies which suggest that a lower dose of PPI may be equally effective in patients who are older or are deficient in the CYP2C19 liver enzyme (3% of whites and African Americans and 17–25% of Asians). Only one of these studies was a head-to-head. comparison, omeprazole versus lansoprazole, but no difference was found between the two.
• While there may be differing effects of the PPIs based on demographics, there are inadequate data to identify any difference between them.

Detailed Assessment