Seropositivity for H. pylori was found in 98 (71%) of 138 symptomatic patients and 70 (51%) of 138 asymptomatic participants (odds ratio [OR], 2.4; 95% confidence interval [CI], 1.4 to 4.0; p<0.001). In the asymptomatic participants, the prevalence of H. pylori infection increased with age, from 18% in the 0- to 10-year age group to 48% in the 11- to 20-year age group, peaking (68%) in the 31- to 40-year age group. In the dyspeptic patients, the prevalence of H. pylori infection was 60% to 73% in all age groups (Table 1). Among persons <21 years old, H. pylori infection was more prevalent in those with symptoms than those without (17 [71%] of 24 vs.12 [38%] of 31; OR, 4.1; 95% CI, 1.1 to 14.9; p=0.02). Similarly, H. pylori seropositivity showed a significant association with dyspepsia among persons 21–30 years of age (35 [73%] of 48 vs. 36 [48%] of 74; OR, 2.6; 95% CI 1.2 to 6.7; p=0.01), but not among persons >30 years of age (46 [70%] of 66 vs. 20 [63%] of 32; OR, 1.4; 95% CI, 0.4 to 2.8; p=0.8).

On bivariate analysis, infection with H. pylori, older age, female sex, working as a manual laborer (>21 years of age), less education, and larger family size (>7 siblings) were associated with dyspepsia in adults (Table 2). H. pylori infection was associated with dyspepsia after adjusting for age, sex, and urban residence (OR, 2.0; CI, 1.1 to 3.3; p=0.02), and among adults, after adjusting for these factors and for education, family size, and occupation (OR, 2.4; 95% CI, 1.1 to 4.9; p=0.02) (Table 3).

This H. pylori serologic study in residents of the Nakuru District of Kenya found the expected relationship between H. pylori prevalence and age among asymptomatic participants. However, among persons with dyspepsia, the prevalence was consistently high for all ages, which yielded an unequivocal association between H. pylori infection and dyspepsia among persons <30 years of age.

Other studies of Africans with dyspepsia have yielded a mean prevalence of 65% (range 60% to 71%), which is consistent with our results (12,13). Our recruitment strategy was very similar to the strategy of a study conducted in Cape Town, South Africa (13). In that 1993 study, H. pylori prevalence among a subset of Africans of non-Caucasian descent with nonulcer dyspepsia attending a gastroenterology clinic was 71%, the same as in our study. However, since the South African study did not include healthy controls, no generalizations can be made about the association between H. pylori and dyspepsia in different part of the continent. Recently, healthy Nigerian adults and dyspeptic patients were found to have similar prevalence of H. pylori infection (80% vs. 88%), but the sample size (50 persons) may have been too small to detect the moderate effects found in our and other’s studies, particularly in the subgroup of persons <30 years of age (14).

The role of H. pylori in dyspepsia is poorly understood (15,16). Dyspeptic symptoms are common in sub-Saharan Africa (17); in some regions, they may account for up to 10% of hospital admissions (18). Because healthcare resources in Kenya are limited, physicians direct diagnostic tests for patients in whom a definitive diagnosis is important for treatment (e.g., those with peptic ulcer or gastric cancer). Since a large fraction of the dyspepsia in younger Africans is attributable to H. pylori, and since dyspepsia in this age group is likely to represent a benign process, a test-and-treat strategy may be appropriate in this age group. This approach involves H. pylori testing of uninvestigated dyspeptic patients without severe symptoms or signs suggestive of underlying malignancy (unexplained recent weight loss, dysphagia, hematemesis or melena, anemia, previous gastric surgery, and palpable mass). Those with positive results would undergo H. pylori eradication therapy before endoscopy is considered. Those testing negative would undergo endoscopy only if dietary and behavioral maneuvers do not ameliorate the complaints.

Our study has several limitations. We did not investigate the underlying causes of dyspepsia, so we may have included an unknown number of participants with peptic ulcer disease or other organic pathology. Additional limitations include the use of convenience (self-selected) controls as a proxy for population controls and the use of prevalence ORs (to make the crude and adjusted risks commensurate) instead of risk ratios. These factors may have led us to overestimate the association between H. pylori and dyspepsia. On the other hand, the use of a test with imperfect sensitivity and specificity may have led us to underestimate this association.

In a recent, randomized, placebo-controlled trial in a developed country, eradication therapy proved successful in a subset of patients with nonulcer dyspepsia (19). However, these findings were not confirmed in another trial of similar design (20). This disparity suggests either that the relationship between H. pylori and nonulcer dyspepsia is weak or that dyspepsia is a heterogeneous disorder. Thus, the effectiveness of a test-and-treat strategy in the developing world may vary by the population studied or by biological and cultural differences in the definition of dyspepsia.

This study demonstrates that in Nakuru, Kenya, H. pylori infection is associated with dyspepsia, particularly in persons <30 years of age. Since solid evidence exists that H. pylori eradication prevents the development (21) and recurrence (22) of gastric carcinoma and promotes regression of B-cell lymphoma of the mucosa-associated lymphoid tissue (MALT) tissue of the stomach (23), the proposed test-and-treat strategy may be an efficient use of health resources in Kenya and perhaps other African countries.