What is the long-term outcome of the different subgroups of functional dyspepsia?
Dr. M. Heikkinen, Department of Medicine, Unit of Gastroenterology, Kuopio University Hospital, Fin-70210 Kuopio, Finland.
Background : Functional dyspepsia is often a long-lasting disorder that accounts for substantial healthcare costs. It has been classified into subgroups assuming that it can guide management of dyspepsia.
Aim : To evaluate the clinical significance of subgrouping functional dyspepsia in a long-term perspective study.
Methods : Consecutive patients with dyspepsia identified by general practitioners were investigated. Those patients with functional dyspepsia (n = 201) were enrolled in this study. Initially, patients were divided into five subgroups (ulcer-like, dysmotility-like, reflux-like, unspecified, and irritable bowel syndrome-like). Patients' medical histories were reviewed after 6–7 years, and the number and outcome of repeated investigations were analysed. At the end of follow-up, patients filled in a questionnaire similar to that at baseline, and were invited for gastroscopy.
Results : Only 2% of patients developed peptic ulcer during follow-up, none of them were in the ulcer-like subgroup. When referrals to hospital and examinations during follow-up were registered, no statistically significant differences existed between subgroups. Patients with reflux-like dyspepsia made fewer revisits than others (P = 0.02), but had used antidyspepsia drugs during the previous year more often (P = 0.036). Stability of the subgroups over time was poor.
Conclusions : Functional dyspepsia is a long-lasting disorder with a very good prognosis. Subgroups of functional dyspepsia play only a minor role in prediction of the long-term outcome, and their usefulness in clinical practice is also hampered by subgroup instability over time.
Dyspepsia diagnosed by general practitioners (GPs) is based mainly on symptoms and on clinical examination. However, diagnosis based only on clinical judgement is unreliable,1–3 and upper endoscopy often must be performed for definite diagnosis. However, as much as 57 to 65% of upper endoscopies done for dyspepsia in primary health care are normal.4, 5 With a normal endoscopy finding, a diagnosis of functional dyspepsia (nonulcer dyspepsia) can be determined. Functional dyspepsia has been classified into subgroups, assuming that this can guide therapy and even explain the pathophysiology of functional dyspepsia.6 However, in population-based studies, although the value of symptom-based subgrouping has been small,7, 8 subdividing of dyspepsia has become entrenched in clinical practice.9
Whether or not the subgrouping plays any role when GPs are considering further investigations or therapy for a patient with persistent or recurrent symptoms is unknown, as well as the predictive value of dyspepsia subgroups or long-term prognosis. The aim of the present study was to assess the role of subgroups of functional dyspepsia in terms of long-term prognosis, need for further investigation and revisits during follow-up, and to guide the use of medication. Secondly, the stability of symptom-based subgroups of functional dyspepsia in carefully investigated groups of patients was analysed.
Initially, 400 consecutive unselected dyspeptic patients were investigated by general practitioners (GPs) in four health centres between January 1993 and January 1994. Their average age was 56.8 years (range 15–86), and 62% were women. At baseline, the definition of Colin-Jones et al. was used for dyspepsia.6 The patients included in the study had abdominal or retrosternal pain, discomfort, heartburn, nausea, vomiting, or other symptoms considered to be referable to the proximal alimentary tract. All patients with no evidence of organic upper abdominal disease explaining their symptoms were diagnosed as having functional disorder and were asked to participate in this follow-up study. However, patients with reflux symptoms only, without upper abdominal complaints, were excluded in accordance with the definition of dyspepsia of Talley et al.10
The study design appears in Table 1. Eligibility criteria were fulfilled by 205 patients. Four patients were lost to follow-up: they lived out of the study area, had no known address, or were unwilling to participate.
Table 1. Classification of functional dyspepsia according to Talley et al.10
|I||Ulcer-like functional dyspepsia|
|Three or more of the following are suggested as necessary, but upper abdominal pain must be a predominant complaint:|
| ||1 Pain that is well localized in the epigastrium (i.e. can be localized to a single small area by pointing with one or two fingers)|
| ||2 Pain relieved by food, often (more than 25% of the time)|
| ||3 Pain relieved by antacids and/or antisecretory drugs, often|
| ||4 Pain occurring before meals or when hungry, often|
| ||5 Pain that at times wakens the patient from sleep|
| ||6 Periodic pain with remissions and relapses (periods of at least two weeks with no pain interspersed with periods of weeks to months when there is pain)|
|II||Dysmotility-like functional dyspepsia|
|Pain is not a dominant symptom. Upper abdominal discomfort should be present in all cases. It is suggested that this discomfort should be chronic and characterized by three or more of the following:|
|4||Retching and/or vomiting that is recurrent|
|5||Bloating in the upper abdomen not accompanied by visible distension|
|6||Upper abdominal discomfort often aggravated by food|
|III||Unspecified functional dyspepsia:|
|This refers to dyspeptic patients whose symptoms do not fulfil the criteria for ulcer-like or dysmotility-like dyspepsia|
At baseline, all patients, regardless of severity of symptoms, underwent gastroscopy, upper abdominal ultrasound, and laboratory tests (B-count, S-ALT, S-alkaline phosphatase, S-amylase, C-reactive protein, and a test for lactose absorption). No selection criteria were used for gastroscopy, for example, Helicobacter pylori tests or empirical drug treatment. Gastroscopies with biopsies were performed by the same specialist, who interviewed the patients thoroughly and after complete investigations settled on the final diagnoses. Upper abdominal ultrasound was performed by an experienced radiologist. Moreover, patients with heartburn or regurgitation as a predominant symptom without endoscopic evidence of oesophagitis underwent a 24-h pH-recording; a pH of below 4 during at least 4% of the recording time was defined as the cut-off level for a pathological result.
Assessment of H. pylori infection
Patients were classified as H. pylori positive or H. pylori negative based on gastric histological specimens or H. pylori serology, or both. Serum samples were tested without knowledge of the patients' diagnoses for IgG antibodies of H. pylori by the enzyme-linked immunoabsorbent assay method (HP CHECK EIA, Hoffmann La Roche ) using a second-generation antigen mixture of external membrane proteins enriched with urease. Sensitivity and specificity of this method has been reported as 99.3 ± 1.3% and 86.5 ± 5.2%.11 In all cases, Giemsa stain allowed detection of H. pylori, but no immunostains were performed. The specimens were reviewed alternatively by two pathologists familiar with gastrointestinal pathology.
Classification into subgroups of functional dyspepsia
Patients were initially asked to fill in a modified version of the Bowel Disease Questionnaire,12 with symptom questions added, classifying dyspeptic patients into subgroups.10 Functional (idiopathic) dyspepsia was divided into ulcer-like, dysmotility-like, and unspecified functional dyspepsia.10 Patients with heartburn or regurgitation as a predominant symptom but with no objective findings of reflux disease were categorized as having reflux-like dyspepsia. Those patients who complained of dyspeptic symptoms and additionally suffered from symptoms characteristic of irritable bowel syndrome (IBS)13 were diagnosed as having IBS-like dyspepsia. All patients classified with reflux-like dyspepsia or IBS had dyspepsia as the main reason for their visit to the GP. The long-term role of subgroup at baseline was evaluated.
The follow-up period ended in September 2000. Patients' medical records were reviewed in their health centres, in Kuopio University Hospital, in local hospitals, and in private clinics as well. The main end-points registered were: diagnosed ulcers, oesophagitis, gastrointestinal malignancies, or other gastrointestinal disease which could have explained symptoms; abdominal operations were also recorded. The total number of gastroscopies, lower abdominal endoscopies, abdominal ultrasounds, and other radiological examinations including abdominal CT were recorded. The time-period for the first revisit, also considered as an end-point, was recorded. Referrals to hospitals were also recorded. Use of drugs for upper abdominal symptoms (proton pump-inhibitors, H2-blockers, antacids, sucralfate, prokinetics) during the previous year was recorded retrospectively.
At the end of the follow-up, all survivor study patients were asked to undergo gastroscopy. Of the study patients who were alive 133 (74%) were thus investigated by one of the authors (M.H.). Any new ulcer scars discovered were recorded as peptic ulcers. Before they were gastroscoped, patients filled in a questionnaire similar to the one at baseline. Others received postal questionnaires. Excluding patients who had died, 89% returned the questionnaire.
All calculations were done with the SPSS 9.0 for Windows program. The Chi-square test was used to analyse the statistical significance of the differences of frequency estimates of categorized variables. Each subgroup was compared with all others. Student's t-test was used to assess the differences in continuous variables between the groups, with 5% considered to significant.
The Medical Ethics Committee of Kuopio University Hospital approved the study.
Visits to GPs
No significant differences existed between mean age, gender, or H. pylori status of the subgroups (Table 2). At the end of follow-up, only 19.4% were asymptomatic. The proportion of symptom-free patients ranged from 14.3% (dysmotility-like) to 25% (ulcer-like). Table 3 shows the proportion of patients in each subgroup who consulted their physicians because of abdominal complaints after the final diagnosis. Patients with reflux-like dyspepsia at baseline made fewer revisits than did other patients (P = 0.02). In contrast, patients with dysmotility-like dyspepsia consulted their physicians more often than did other patients with dyspepsia, but the difference was not statistically significant (P = 0.06). Mean time for the first revisit was 34.9 (S.E.M.; 5.3) months for ulcer-like, 24.4 (4.3) for dysmotility-like, 26.3 (3.7) for unspecified, 33.6 (7.5) for reflux-like, and 31.7 (5.3) for IBS-like dyspepsia subgroups, with differences not significant.
Table 2. Demographics of patients in subgroups of functional dyspepsia during follow-up
|Age years (S.E.M.)||56.1 (2.47)||56.9 (2.65)||58.2 (2.10)||58.8 (2.70)||52.2 (2.32)|
|H. pylori-positive||17 (55%)||19 (51%)||42 (66%)||21 (66%)||19 (51%)|
|Survivors asymptomatic study end*||7 (25%)†||5 (14%)||12 (21%)‡||6 (21%)§||5 (14%)†|
Table 3. Proportion of patients in each subgroup who had revisits, referrals to gastroscopy, to sigmoideo- or colonoscopy, to abdominal ultrasound during follow-up, and who used medication for their abdominal complaints during the previous year
|Sigmoideo- or colonoscopy||32%||22%||25%||16%||41%†|
|Use of drugs for abdominal complaints§||36%||31%||28%||37%‡||21%|
Further investigations and findings
Diagnostic procedures performed during follow-up are shown in Table 3. The number investigated by endoscopy (gastro-, sigmoideo- or colonoscopy) was 79 of 201 patients (39%). During the follow-up period, patients with IBS-like dyspepsia at baseline underwent sigmoideo- or colonoscopy more often than did other patients (41% vs. 24%; P = 0.038). The subgroups of functional dyspepsia showed no significant differences between the number on whom gastroscopies or abdominal ultrasounds had been performed. About one-fifth of the study patients had undergone another radiological procedure (e.g. computerised tomography scan) with no differences between subgroups. Of all patients, 16.9% were referred to hospital investigations during follow-up (ulcer-like, 16.1%; dysmotility-like, 16.2%; unspecified, 12.5%; reflux-like, 15.6%; IBS-like, 27%), but with no significant differences when each subgroup was compared to all other patients. Organic gastrointestinal diseases diagnosed in the follow-up period are presented in Table 4. No differences appeared between the study groups in the proportion of patients who developed organic gastrointestinal problems. Four peptic ulcer diseases were diagnosed, and all of these patients were H. pylori-positive. Of patients who developed oesophagitis, only two had reflux-like dyspepsia at baseline. Altogether eight patients underwent surgery for gallstone disease, and one had acalculous cholecystitis treated by surgery. Colorectal carcinoma was found in colonoscopy performed on a 61-year-old-female who had already undergone surveillance colonoscopy because of small adenomas prior to initial investigations for dyspepsia. During follow-up, the flat polyp (diameter 1.5 cm) found turned out to be malignant. Surgical treatment was curative. A 72-year-old-female died of diverticular perforation a year after final diagnosis. No-one else amongst study patients had gastrointestinal malignancies, nor any fatal events related to gastrointestinal diseases.
Table 4. Organic gastrointestinal diseases in patients in subgroups of functional dyspepsia during follow-up
|Number of diseases||2 (7%)||6 (16%)||5 (8%)||4 (13%)||7 (19%)|
| Peptic ulcer disease||0||0||1||1||2|
| Cholecystectomies done||0||3||1||1||4|
Drugs for upper abdominal complaints were used during the previous year by 54 of 180 study patients (30%) alive at the end of follow-up. Patients with reflux-like dyspepsia had used drug treatment (antacids, sucralfate, H2-blockers, proton pump inhibitors, or prokinetics) more often (37% vs. 29%) than had other patients with dyspepsia (P = 0.036). There were no significant differences between other subgroups (Table 3).
Stability of subgroups over time
Stability of the subgroups over time is presented in Table 5. After excluding cases in which the subgroup was unknown at the end of follow-up, 37% of patients with ulcer-like, 20% of dysmotility-like, 22% of unspecified, 13% of reflux-like, and 34% of IBS-like dyspepsia had same subgroup as at baseline.
Table 5. Stability of subgroups at the end of follow-up
This may be the first long-term follow-up study of the natural history of a carefully investigated group of patients with functional dyspepsia and the impact of symptom subgroups as a prognostic factor. In contrast to most earlier studies, our results were not based solely on a questionnaire. In addition, patients' medical records were evaluated, and upper endoscopies were performed on 74% of the study patients at the end of the follow-up.
In dividing our study patients into subgroups of functional dyspepsia, we were aware of the great overlap between dyspepsia and gastro-oesophageal reflux disease (GERD) and IBS.7, 8, 14, 15 However, in ‘real life’, patients consulting their GPs for upper abdominal complaints, as our study patients did, are probably managed like those dyspeptic patients who do not have concomitant IBS- or reflux-like symptoms. Reflux disease has been suggested to be separate from dyspepsia.10 In our study, all patients with heartburn or regurgitation only were excluded, and those with a predominant reflux-like syndrome who had a pathological result in a 24 h pH measurement were excluded, as well. The remaining patients with a predominant reflux syndrome were classified as having reflux-like functional dyspepsia. These patients also had other dyspeptic symptoms like upper abdominal pain, bloating, or early satiety. However, some might have GERD, because pH monitoring is not 100% sensitive.16–19 The most controversial subgroup of the present study was IBS-like functional dyspepsia. Despite the latest suggestions,10, 19 we did not exclude these patients, because all had dyspeptic symptoms as their main reason to visit a physician. In fact, these patients had an unpleasant or troublesome, painful or nonpainful sensation centred in the upper abdomen and thus could have been classified as having ulcer-like or dysmotility-like functional dyspepsia.19
Symptoms and visits
Only one-fifth of the patients who were alive at the end of follow-up were asymptomatic, subgroup dividing did not play a major role in prediction of disappearance of symptoms in the long-term. Recently, Agréus et al. have reported in their 7-year follow-up based only on a symptoms questionnaire that 40% of individuals with dyspepsia, 34% with IBS, and 61% with predominant reflux symptoms became symptom-free or had only minor symptoms.20 In other long-term follow-up studies,21–24 18–36% of individuals with dyspepsia were free of symptoms after 3 to 14 years of follow-up. In contrast to our study, these studies were based on postal questionnaires20–22, 24 or on uninvestigated dyspepsia.20 Although these studies are epidemiologically important, our study is more relevant from a clinical point of view. The crucial question is: What is the long-term prognosis of those dyspeptic patients who seek medical care, after a negative (normal) upper endoscopy? Patients with reflux-like dyspepsia made revisits less often than did other patients. A probable explanation would be that these patients used medication for their symptoms more often than did other patients. Our finding is in agreement with the results of Meineche-Schmidt and Jørgensen, who have also investigated how patients with different subgroups are managed by GPs.25 In our study, none of the patients in the subgroups differed from other patients with upper abdominal complaints in respect to mean time before the first revisit.
Investigations and organic diseases
During a 6- to 7-year follow-up, quite a high proportion of our study patients underwent repeated investigations. Patients with reflux-like dyspepsia underwent gastroscopy less often than did other dyspeptic patients, one reason for which may be that reflux-predominant symptoms without oesophagitis are considered a benign phenomenon with a clear-cut symptom profile26 or that symptoms are well controlled by use of H2-blocker or proton pump inhibitors. The IBS-like dyspepsia subgroup was the only one which differed in comparison to the group of all other dyspeptic patients regarding the rate of lower gastrointestinal endoscopies performed. The need for these examinations is understandable, because these patients have from time to time also concomitant lower abdominal symptoms. This study, however, confirmed the low diagnostic yield of endoscopies for this subgroup.
The study patients lived in an area in which medical records could be reliably inspected.27 It would have been obvious that complications such as perforations, bleeding, and hospitalization due to serious events including death would had been recorded. Only 10–11% of patients developed an organic gastrointestinal disease. Peptic ulcer disease was found in four cases, all of whom were H. pylori positive. It could be argued that peptic ulcers can heal spontaneously, leading to a higher incidence of peptic ulcer disease than we found. However, 74% of patients who were alive at the end of follow-up underwent voluntary gastroscopy, and all with an ulcer scar were diagnosed as having peptic ulcer disease.
Patients suffering from functional dyspepsia, like those in our study, are at a very low risk of complicated peptic ulcer disease.21–25, 28 Oesophagitis was the most common organic gastrointestinal disease diagnosed during follow-up, and all of these cases were mild (Los Angeles grade A), with typical symptoms. These patients could have been well-healed with empirical drug treatment.26
Patients with reflux-like dyspepsia used medication for their abdominal complaints more often than did others. This is most likely explained by better response to treatment,29 probably leading to the lower revisit rate in this subgroup.
Subgroup stability over time
Our results confirm the findings of Agréus et al. who reported marked instability of dyspepsia subgroups.14 From this point of view, the decision to include reflux-like and IBS-like patients in our study seemed reasonable. In previous studies assessing the value of subgroups comprising patients with uninvestigated dyspepsia. In these studies it was estimated that most of the patients would have functional dyspepsia. In the present long-term follow-up study, 63–87% of patients in different groups changed their subgroups. In contrast to previous reports concerning a proven GERD,30–32 patients with reflux-like dyspepsia also had an unstable subgroup. In our study, however, patients with true GERD were most probably excluded by typical symptoms and by pH recording. The instability of subgroups in functional dyspepsia makes it questionable to use for guiding the management of dyspepsia, at least in the long-term perspective.
A limitation of our study is the relatively low number of study patients, leading to a low number of end-points, although our series seems to be the largest long-term follow-up study on the clinical value of subgroups of investigated functional dyspepsia conducted so far. Moreover, to conduct this kind of follow-up study with a higher number of patients is troublesome. The health care organization in the study area27 provided us with an excellent opportunity to collect reliable information from patients' medical records.
The fact that data on drug usage during the previous year were collected retrospectively is another limitation. This may have an effect on the number of patients reporting that they had been under medication, but this can hardly affect the results of the study: the proportion of drug usage in each subgroup was comparable with that for all other patients.
In conclusion, the subgroups of functional dyspepsia have a low impact on clinical practice, in terms of long-term prognosis and need for endoscopic or radiologic examinations during follow-up. The subgroups may, however, be useful in clinical studies to select certain types of patients for research.