SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

Background : Lacking an objective ‘gold standard’ for diagnosing dyspepsia, several symptom-based classifications have been suggested.

Aim : To assess if response to proton-pump inhibitor treatment could provide useful information for current or future dyspepsia classification.

Methods : Post hoc analyses of 829 patients treated with omeprazole or placebo in a randomized-controlled trial. The ‘true’ response to omeprazole (trial response minus placebo response) was assessed according to different classifications of dyspepsia and different symptoms.

Results : Symptoms described with the words ‘burning’ or ‘sour’ and patients with reflux-like dyspepsia demonstrated high response to omeprazole treatment, whereas patients with abdominal pain or ulcer-like dyspepsia responded unpredictably to omeprazole. The response to omeprazole in patients with epigastic pain was related to the pattern of other dyspeptic symptoms. Patients with heartburn or regurgitation overlapped extensively with patients with epigastric pain.

Conclusion : The study demonstrated significant problems in the current classification of dyspepsia: ‘the most bothersome symptom’ was not independently related to the omeprazole effect and, in patients with abdominal pain, the response to omeprazole was dependent on the presence or absence of other dyspeptic symptoms. The overlap of symptoms indicates that heartburn and regurgitation should be recognized as symptoms of dyspepsia in primary care.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

Classification of patients with dyspeptic symptoms has changed several times over the last decades. Early classification was based on clusters of symptoms among all reported gastrointestinal symptoms.1 Gastro-oesophageal reflux disease (GERD) and irritable bowel syndrome (IBS) were included in the definition of dyspepsia. Later on IBS was regarded as a separate disease and finally GERD too was defined as a separate entity. Dyspepsia was now restricted to patients reporting symptoms from the upper abdomen that was not dominated by heartburn and acid regurgitation. This classification was initially based on ‘clusters of symptoms’, Rome I,2 but in the latest version on ‘the predominant symptom’, Rome II.3 The overall trend in definitions has therefore been reductionism.

While this may have been appropriate in terms of research, it has created problems in the daily clinic, where the patients seem to present with complex rather than simple symptoms.

There is a lack of ‘gold standard’, by which dyspepsia can be classified clinically.

Endoscopy,4 and Helicobacter pylori testing5 has been suggested as gold standard, but these methods are not applicable in primary care, first of all because treatment in primary care often is symptomatic and empiric,6 and secondly, no solid link between endoscopic or H. pylori diagnosis and response to treatment has been demonstrated.7

The vast majority of patients in primary care are treated empirically,6 which is allows the patient to experience the treatment response, which may not be possible if the treatment is based on an endoscopy or a paraclinical test. In primary care, no difference in the treatment of patients with ‘reflux’ symptoms and patients with ‘ulcer’ symptoms have been found6 and a common diagnose: ‘acid-related symptoms’ seems practical in primary care.

Based on a large-scale epidemiological study, comprising more than 7000 patients,8 who consulted in general practice because of ‘any type of abdominal symptom’ it was demonstrated that on the average one patient reported 5.5 symptoms, if a systematic interview was performed. In a randomized-controlled proton-pump inhibitor (PPI) trial patients with ‘predominantly ulcer-like or reflux-like dyspepsia’ reported 5.2 symptoms on the average9 and in another PPI trial patients who consulted in general practice because of ‘acid-related symptoms’ reported 5.2 symptoms on the average.10 This indicates that selection criteria per se act in a qualitative rather than quantitative manner.

With more than five symptoms present in the average patient, it is important to study how the doctors interpret the symptoms presented by the patient. Will the doctor be focused on specific symptoms or on all symptoms? Will some symptoms be ‘highlighted’ and other ‘neglected'?

For research purposes, it is important to be able to identify a patient population. This is done by a priori defined criteria for classification (inclusion into a study for instance) and will often be based on the presence of symptoms that are focused in the study, combined with demographic data. In the clinical situation, the diagnostic process is different. The doctor will often have a theory a priori of the nature of the complaint. In primary care General Practitioners (GPs) use ‘pattern-recognition’, i.e. has something like this been presented to me before? What did I do? And how did it work? Talking with the patient will allow the GP to search for confirmatory as well as contradictory information resulting in either acceptance of the identified pattern (which then becomes a ‘diagnose’) or search for a better one. This is a hermeneutical approach11 and it allows the use of all available information for the diagnostic process in contrast to a reduced set of information presented as inclusion and exclusion criteria.

Furthermore, some patients in the clinic will express symptoms different from those defined a priori and the doctor will have to ‘adapt’ the self-reported symptom into one of the predefined symptoms. The hermeneutical approach11–13 will respect the self-reported symptom, the doctor will discuss to clarify the meaning of this and will take into account the presence of ‘other’ symptoms, before a pattern-recognizing diagnose is established. The consequence should be a better identification of patients leading to an improvement in response to therapy. This is particularly important in primary care, because lack of response to treatment in one indication for diagnostic work-up, which is time-consuming and costly to both patients and society.

In the single patient, it is difficult to distinguish between a favourable response caused by ‘handling’ the patient = placebo response, and a favourable response caused by the prescribed drug. Trials have been designed in which the change from placebo to active drug was blinded, but the results in term of identifying responders have been disappointing.14 In the set-up of a randomized, placebo-controlled trial it is possible to compare group of patients treated with placebo to group of patients treated with active drug, and the difference in response between these groups can be considered as the true effect of the drug. In the present study, the ‘Absolute Risk Reduction’ (ARR) of omeprazole treatment was calculated as the difference between the response to omeprazole and the response to placebo.

By comparing the ARR in different dyspepsia classification or in single dyspeptic symptoms it was possible to assess the ‘acid-related nature’ of these classifications and symptoms. Furthermore, the effect of ‘adaptation’, i.e. the doctor's interpretation of the patient-reported symptoms in the diagnostic process could be addressed by comparing the ARR according to the patients’ self-reported symptom and a priori defined classifications. The validity of dyspepsia classification based on one symptom (the most bothersome) was tested by assessing changes in the ARR in patients with or without other dyspeptic symptoms.

Aim of the study

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References
  • A
    To assess the ARR of omeprazole treatment according to three different dyspepsia classifications: a classification based on the predominant cluster of symptoms, a classification based on ‘the most bothersome symptom (MBS)’ and a classification based on the patient’ own phrasing of the reason for consulting the GP.
  • B
    To compare the ARR of omeprazole treatment in dyspepsia classifications based on ‘MBS’ and ‘the reason for consulting’.
  • C
    To assess the robustness of classification based on ‘MBS’.

Material and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

The analyses in the study were based on the result of a randomized-controlled trial, conducted in 2000–01,10 which can be summarized to: 283 patients were treated with omeprazole 40 mg daily, 273 patients with omeprazole 20 mg daily and 272 patients with placebo. The treatment period was 2 weeks. Inclusion criteria was based on the assumption by the GP, which the symptom that prompted the consultation could be abolished by treatment with an acid-reducing drug, i.e. the symptom was ‘acid-related’. The response to omeprazole 40 mg was 66.4%, omeprazole 20 mg 63.0% and placebo 34.9%. As the response to high- and low-dose of omeprazole was not significantly different, the patients were analysed together.

Dyspeptic symptoms were assessed in the following way: first, the patient was asked to phrase in own words the reason for consulting the GP (the Main Symptom). Secondly, the patient was interviewed about the presence or absence of 10 different dyspepsia symptoms (listed in Table 1) and finally the patient was asked to identify, which of those symptoms present was the ‘most bothersome’.

Table 1.  Classification of dyspepsia according to clusters of symptoms (SCC), the most bothersome symptom (MBSC) and the patient's own phrasing of the reason for consulting the GP (SC) in 829 patients consulting general practice because of acid-related complaints
Dyspepsia symptomSCCMBSCSC
N%ClassN%ClassN%Class
  1. Number and percentage of patients, and defined dyspepsia class: A, ‘acid component’; D, dysmotility-like dyspepsia or ‘discomfort component’; P, ‘pain component’; R, reflux-like dyspepsia; U, ulcer-like dyspepsia; SCC, symptom cluster classification; MBSC, most bothering symptom classification; GP, General Practitioner; SC, self-classification.

Epigastric pain73589U45955U36743P
Heartburn56568R19023R20224A
Regurgitation57970R638R617A
Bloating50161D202D111D
Pain at night44454U425U142P
Belching38647D41D1D
Discomfort after meals38647D142D1D
Early satiety36244D00D
Nausea35343D374D324D
Postprandial fullness15919D00D
Burning in the abdomen789A
‘Cramps’, ‘abdominal stones’465D
Miscellaneous   162D

All patients were classified into three different dyspepsia classifications:

Symptom cluster classification

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

Symptom cluster classification (SCC) was done by interview about the presence or absence of 10 dyspeptic symptoms (Table 1). These symptoms were grouped into three different subgroups and the patient was classified according to the ‘weight’ of these subgroups into: ‘Reflux-like’: if the sum of points of symptoms marked with ‘R’ in Table 1 was higher than the sum of ‘U’ and ‘D’. ‘Ulcer-like’: if the sum of points of symptoms marked with ‘U’ in Table 1 was higher than the sum of ‘R’ and ‘D’ and ‘Dysmotility-like’: if the sum of points of symptoms marked with ‘D’ in Table 1 was higher than the sum of ‘U’ and ‘R’.

Reflux-like and ulcer-like symptoms were weighted three times as heavily as dysmotility-like symptoms, allowing a total score in each subgroup of 6. Any combinations of equal scores were not included in the analyses.

Most bothersome symptom classification

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

Most bothersome symptom classification (MBSC) was done according to the MBS by interview, into ‘Reflux-like’: if any symptom marked ‘R’ in Table 1 was the most bothering, ‘Ulcer-like’: if any symptom marked ‘U’ in Table 1 was the most bothering and ‘Dysmotility-like’: if any symptom marked ‘D’ in Table 1 was the most bothering.

Self-classification

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

Self-classification (SC) was based on the Main Symptom and was classified as ‘Acid component’: if any symptom marked ‘A’ in Table 1 was the reason. ‘Pain component’: if any symptom marked ‘P’ in Table 1 was the reason and ‘Discomfort component’: if any symptom marked ‘D’ in Table 1 was the reason.

The patient phrased Main Symptom was grouped by the author into items, comparable with those in the interview, whenever possible. However, three new classes of symptoms had to be defined, as a total of 140 patients (17%) phrased Main Symptoms that was not comparable with any of the 10 listed symptoms.

To assess the ‘acid-related’ nature of different classifications (Study Aim A) the ARR of omeprazole treatment was calculated in group of patients fulfilling the definitions of the above defined classes, to assess the effect of ‘interpretation’ of patient phrased symptoms (Study Aim B) the ARR was compared between MBSC and SC and to assess the robustness of MBSC (Study Aim C) the ARR in MBSC classes were compared according to the presence or absence of additional dyspeptic symptoms.

Comparison between the ‘acid-related’ nature of different classifications was tested by 95% confidence intervals of the ARR of omeprazole treatment.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

About 829 patients reported a total of 4470 symptoms, i.e. 5.4 symptoms in average. Allocation of the reported symptoms to dyspepsia classes is given in Table 1. Complete data, allowing evaluation of response to active drug and placebo was obtained in 820 patients.

Symptom presentation

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

Epigastric pain was the most common symptom reported in all classifications. In SCC, all symptoms present took part in the classification and heartburn and regurgitation were very common, but ‘dysmotility-like’ symptoms were frequent and even the least reported symptom (postprandial fullness) was present in one of five patients. MBSC and SC classification allowed one symptoms only for classification and the frequencies could be compared: small differences were found for all symptoms except epigastric pain (55% vs. 43%) and pain at night (5% vs. 2%).

Study aim A

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

Figures 1–3 show the allocation into different dyspepsia classes, the response to active drug and placebo, and the ARR of omeprazole treatment with 95% CI. In the SCC (Figure 1), the ARR in patients with dysmotility-like dyspepsia was not significantly different from the ARR in patients with reflux-like dyspepsia and significantly higher than in patients with ulcer-like dyspepsia. Patients with reflux-like dyspepsia had a significantly higher ARR compared with patients with ulcer-like dyspepsia. In the MBSC (Figure 2), the ARR in patients with dysmotility-like dyspepsia was neither significantly different from the ARR in patients with reflux-like dyspepsia nor patients with ulcer-like dyspepsia. Classified according to SC (Figure 3), the patients reporting ‘discomfort component’ demonstrated a significantly lower ARR than patients reporting ‘acid component’, and a significantly higher ARR than patients reporting ‘pain component’. The ARR in patients with ‘acid component’ was significantly higher than the ARR in patients with ‘pain component’.

image

Figure 1. Symptom cluster classification: dyspepsia classification according to weight of symptom clusters. Number of patients treated with active drug and with placebo in each dyspepsia class, the effect of treatment and the absolute risk reduction (ARR) (with 95% confidence interval, CI).

Download figure to PowerPoint

image

Figure 2. Most bothersome symptom classification: dyspepsia classification according to the most bothersome symptom by interview. Number of patients treated with active drug and with placebo in each dyspepsia class, the effect of treatment and the absolute risk reduction (ARR) (with 95% confidence interval, CI).

Download figure to PowerPoint

image

Figure 3. Self-classification: dyspepsia classification according to the patient's own phrasing of the reason for consulting. Number of patients treated with active drug and with placebo in each dyspepsia class, the effect of treatment and the absolute risk reduction (ARR) (with 95% confidence interval, CI).

Download figure to PowerPoint

In all classifications, the ARR were highest in the reflux-like dyspepsia/‘acid component’: 45% in the SCC, 40% in the MBSC and 43% in SC. The comparable figures in the dysmotility-like dyspepsia/‘discomfort component’ were 45, 35 and 30% and in the ulcer-like dyspepsia/‘pain component’ 18, 25 and 19%.

Study aim B

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

A total of 140 patients (17%) phrased the reason for consulting in a way (belly cramps, burning belly, stones in the abdomen) that could not be classified according to the symptoms listed in the interview (Table 1). Most of these patients were classified as epigastric pain in MBSC (Table 2). Especially self-reported burning sensation in the abdomen tended to be classifies as epigastric pain. The MBSC was unable to demonstrate a significantly different ARR in patients with ulcer-like dyspepsia compared to patients with dysmotility-like dyspepsia, which was in contrast to the SC.

Table 2.  The relation between the patient's own phrasing of the reason for consulting the GP (SC) and most bothersome symptom by interview (MBSC) in 829 patients consulting in general practice because of acid-related complaints
SC classification*MBSC classification†
Heart burnRegurgitationPostprandial fullnessBloatingNauseaPain at nightEpigastric pain
  1. MBSC, most bothering symptom classification; GP, General Practitioner; SC, self-classification.

  2. * Two patients with belching and one patient with discomfort after meal excluded.

  3. † Four patients with belching and one patient with discomfort after meal excluded.

Burning in the throat (n = 202)159213334
Sour sensation in the mouth (n = 60)743217
Air in the abdomen, balloon sensation (n = 11)812
Nausea (n = 32)22271
Disturbed sleep due to abdominal pain (n = 14)14
Pain in the upper part of the abdomen (n = 364)5863414324
Abdominal cramps, sense of abdominal ‘stones’ (n = 46)42362425
Burning sensation in the abdomen, ‘flames’ (n = 78)1252356
Other complaints (n = 16)311110

Study aim C

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

Other dyspeptic symptoms influenced the ARR of omeprazole treatment in patients, who reported heartburn or regurgitation as the MBS (Table 3): the presence of early satiety, postprandial fullness, nausea or nocturnal pain decreased the ARR, compared to patients without these symptoms, whereas the presence of bloating and ructations increased the ARR. Comparable results were found in patients classified with epigastric pain as MBS (results not shown). The differences were related to changes in the placebo response (primarily) as well as the response to active drug.

Table 3.  The effect of active treatment and placebo and the true effect of omeprazole (ARR) according to the presence or absence of other dyspeptic symptoms in 253 patients who reported heartburn or regurgitation as the most bothersome symptom when consulting general practice because of acid-related complaints
 N*Active treatment (%)Placebo (%)ARR (%)Difference ARR (95% CI)
  1. ARR, absolute risk reduction; CI, confidence interval.

Early satiety
 Absent16769264312 (−1 to 24)
 Present82744331 
Postprandial fullness
 Absent15375245129 (17–39)
 Present96644222 
Bloating
 Absent113734132−14 (−26 to 2)
 Present136692346 
Ructation
 Absent132743639−4 (−15 to 9)
 Present117672443 
Nausea
 Absent15876235742 (30–50)
 Present91634815 
Pain at night
 Absent1517834449 (−2 to 22)
 Present98612635 

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

Omeprazole blocks the proton pump and no other effects of the drug are known. The basic idea behind the study was to interpret a beneficial effect of the drug as an indication of a link between the symptom and acid in the gastrointestinal tract. These post-trial analyses were done in order to get new ideas for classification of patients with dyspeptic complaints, based on the true response to an acid-blocking agent. The analyses are of course restricted to the study population and must be confirmed in other patient populations.

It is strength in the study, which the ‘true’ omeprazole response was used to assess the relation between symptoms, classifications and acid, because placebo responses varied greatly. The overall result of the RCT, which formed the basis for this study, was high: 64% and the placebo effect was moderate: 34%, compared with other placebo-controlled PPI trials,9, 15, 16 which meant that one patient in three diagnosed as having ‘acid-related’ symptoms was beneficially treated by omeprazole.

The study was conducted in a way that made it possible to classify all participating patients according to different classification, one based on clusters of symptoms (which formed the basis for the Rome I classification) and two based on a single symptom, either the most bothering (comparable with the Rome II classification) or the reason for consulting the GP. First, the patient's own phrasing was obtained and later on the patient was interviewed about the presence or absence of 10 dyspepsia symptoms – and finally the patient identified the MBS. This way of taking the history was chosen because we did not want to ‘induce’ symptoms into the patient but at the same time we wanted to record all symptoms.

Study aim A

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

The lack of ability of the SCC to demonstrate a greater true effect of omeprazole in patients with reflux-like dyspepsia compared with dysmotility-like dyspepsia questions the validity of SCC, if all symptoms are equally weighted. However, almost one-third of the patients were excluded from the analysis, because of equal scores in more than one subgroup, leaving only very few patients for analysis in the dysmotility class. The MBSC also failed to demonstrate a significantly better true effect of omeprazole in both reflux-like and ulcer-like dyspepsia compared with dysmotility-like, which question the validity of classifying by one symptom only. The SC classification demonstrated a significantly higher effect of omeprazole in patients with acid compared to patients with discomfort component, which on the other hand, demonstrated a significantly higher effect than patients with pain component. In all classifications, patients with reflux-like/‘acid component’ demonstrated a highly significantly better response to omeprazole compared to patients with ulcer-like dyspepsia/‘pain component’. While symptoms reported by using the word ‘burning’ had a uniformly high ARR (40–45%), epigastric pain demonstrates an unpredictable ARR (18–25%). Probably, pain could be caused by different aetiologies: acid or distension (because of motility disturbances?). The pain localization has been used for discriminating acid pain (located in the epigastric region) and distension pain (located in the abdomen), but this might be questioned. Another way of discriminating pain could be to look at ‘additional’ symptoms. ‘Epigastric pain relieved by food’ could indicate aetiology different from ‘epigastric pain accentuated by eating’. The classic ulcer pain was described as located in the top of the abdomen, and with a favourable response to food or antacids, which is in contrast to epigastric pain accompanied by ‘early satiety’ or ‘unpleasant feeling after meals’.

Testing the idea of different aetiologies for epigastric pain we hypothesized that epigastric pain in patients with both ‘postprandial fullness’ and ‘early satiety’ as additional symptoms (n = 119) was caused by factors other than acid and we revised the SC classification by shifting these patient from ‘pain component’ to ‘discomfort component’. The ARR of Omeprazole treatment was markedly changed in a clinical meaningful direction: patients with ‘acid component’ had a significantly higher response to omeprazole, compared to patients with ‘pain component’, which again had a significantly higher response than patients with ‘discomfort component’ (Figure 4).

image

Figure 4. Revised self-classification: dyspepsia classification according to the patient's own phrasing of the reason for consulting. Patients with epigastric pain + early satiety + postprandial fullness classified as ‘discomfort’ component. Number of patients treated with active drug and with placebo in each dyspepsia class, the effect of treatment and the absolute risk reduction (ARR) (with 95% confidence interval, CI).

Download figure to PowerPoint

Study aim B

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

Comparison between the patients own phrasing of the reason for encounter and ‘most bothersome’ symptom by interview was done in order to address the interpretive effect of the patient–doctor contact leading to a clinical diagnose. This aspect is particularly important, when MBSC is used, because it depends on one symptom only. In the majority of the patients no difference was observed between these methods of classifying, but a total of 140 patients (17%) phrased the reason for encounter in a way that was not comparable with symptoms in the interview list. The majority of patients reporting ‘burning sensation’ in the abdomen as well as the majority of patients reporting ‘sensation of stones’ in the abdomen were classified as ‘epigastric pain’ in MBSC (Table 2). This indicates that when ‘epigastric pain’ is present, it might represent a variety of self-experienced symptoms. This is an important problem in research, because inclusion criteria for trials are normally one or more predefined symptom(s). Classification of patients by scoring systems, or sheets may thus lead to incorrect interpretations of the symptoms and incorrect inclusion.

Study aim C

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

Although the patients in the trial were selected as having acid-related disturbance, high numbers reported symptoms indicating motility disturbance as well (Table 1). We tested if the MBSC was independent of the presence or absence of such additional symptoms. The presence of symptoms indicating motility disturbances had a crucial effect on the ARR of omeprazole treatment in patients classified as having reflux-like or ulcer-like dyspepsia according to MBSC, i.e. the presence of nausea in patients with reflux-like dyspepsia decreased the ARR from 57 to 15% (Table 3). This indicates that MBSC, based on one single symptom could not predict response to PPI even in patients with reflux-like dyspepsia and the validity of the classification must be questioned.

Is there ‘another way’ to Rome?

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

The overall result of this study was, which the ARR of omeprazole treatment was high in patients, who reported symptoms by using the words ‘acid’, ‘burning’ or ‘sour’, whether or not the location was behind the sternum or in the abdomen. On the contrary, patients reporting pain or discomfort had unpredictable ARR. The Rome I criteria were based on the assumption that all symptoms were equally important, whereas the Rome II criteria were based on the opposite: all symptoms except the most bothersome could be neglected in the diagnostic process.

Maybe a ‘way in between’ could be helpful? Patients reporting, ‘burning sensation’ in the abdomen as a bothering symptom could be considered having an acid-related disturbance. Patients reporting ‘epigastic pain’ as a bothering symptom should be interviewed about additional symptoms. If the pain is ‘relieved by food or antacids’, an acid-related disturbance should be considered. If, on the other hand, additional symptoms as ‘nausea’, ‘postprandial fullness’ or ‘early satiety’ are present a non-acid-related disturbance should be considered. The practical importance is to ask for all symptoms, but not necessarily judge all symptoms equally. It is necessary to address not only the presence but also the significance and the patient’ s interpretation of the symptoms. The talk with the patient should be continued until the symptom pattern cannot be improved, i.e. better understood.

Based on the fact that most patients with dyspepsia are treated in primary care and treated empirically,6 it seems reasonable to base classification on symptoms rather than endoscopy or serology. Symptom-based classification, however, has to reflect the fact that patients reports more than five symptoms, on average. The attempt to overcome this problem by concentrating on the MBS seems to be flawed by the fact that the effect of acid reducing treatment is depending on the pattern of symptoms rather than the nature of the MBS.

No doubt, the Rome II classification with focus on one symptom only, has been a major reason for excluding ‘reflux disease’ from dyspepsia. However, this study demonstrates that the majority of patients report heartburn and epigastric pain at the same time. The ‘true’ GERD patient with heartburn and/or regurgitation but without epigastic pain was a rare bird – in this study <10% of the patients presented with ‘true’ GERD.

From a primary care point-of-view, epigastric pain, heartburn and regurgitation as well as burning sensations in the abdomen should be considered symptoms of acid-related dyspepsia, which means that the definition of dyspepsia should include reflux symptoms. Lately, this approach has been advocated by other researchers.17

Conclusion

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References

In this post-trial-analysis, the ‘true’ response of the PPI omeprazole did not reflect the current dyspepsia classification.

The response was related to the pattern of symptoms rather then the most bothering symptom. Patients who described their symptoms as ‘burning’ had a beneficial effect of omeprazole, whether or not the symptoms were referred to the upper or lower abdomen. Patients who described their symptoms as ‘pain’ had an unpredictable effect of omeprazole. In patients with abdominal pain, analysis of additional dyspeptic symptoms may help identifying the symptom pattern that responds favourable to acid inhibition.

Most patients with GERD symptoms reported at the same time epigastric pain and a diagnostic separation between dyspepsia and GERD may not be appropriate in primary care.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Aim of the study
  5. Material and methods
  6. Symptom cluster classification
  7. Most bothersome symptom classification
  8. Self-classification
  9. Results
  10. Symptom presentation
  11. Study aim A
  12. Study aim B
  13. Study aim C
  14. Discussion
  15. Study aim A
  16. Study aim B
  17. Study aim C
  18. Is there ‘another way’ to Rome?
  19. Conclusion
  20. Acknowledgement
  21. References
  • 1
    Colin-Jones DG, Bloom B, Bodemar G. Management of dyspepsia: report of a working party. Lancet 1988; i: 5769.
  • 2
    Talley NJ, Colin-Jones D, Koch KL, et al. Functional dyspepsia: a classification with guidelines for diagnosis and management. Gastroenterol Int 1991; 4: 14560.
  • 3
    Talley NJ, Stanghellini V, Heading RC, et al. Functional gastroduodenal disorders. Gut 1999; 45: 113742.
  • 4
    Bytzer P, Hansen JM, Schaffelizky de Muckadell OB. Empirical H2-blocker therapy or prompt endoscopy in management of dyspepsia. Lancet 1994; 343: 8116.
  • 5
    McColl KE, El-Nujumi A, El-Omar E, Dickson A. Symptomatic benefit 1–3 years after H. pylori eradication in patients. Am J Gastroenterol 2000; 95: 1015.
  • 6
    Meineche-Schmidt V, Jorgensen T. Investigation and therapy in patients with different types of dyspepsia: a 3 year follow-up study from general practice. Fam Pract 2000; 17: 51421.
  • 7
    Hansen JM, Bytzer P, Scaffelizky de Muckadell OB. Management of dyspeptic patients in primary care. Scan J Gastroenterol 1998; 33: 799805.
  • 8
    Meineche-Schmidt V, Krag E. Dyspepsia in general practice in Denmark. A 1-year analysis of consulters in general practice. Scan J Prim Health Care 1998; 16: 21621.
  • 9
    Meineche-Schmidt V, Krag E. Antisecretory therapy in 1017 patients with ulcerlike or refluxlike dyspepsia in general practice. Eur J Gen Pract 1997; 3: 12530.
  • 10
    Meineche-Schmidt V. Empiric treatment with high and standard dose of omeprazole in general practice. Two-week randomised placebo-controlled trial an 12-months follow up of health care consumption. Am J Gastroenterol 2004; 99: 10508.
    Direct Link:
  • 11
    Svenaeus F. The Hermeneutics of Medicine and the Phenomenology of Health. Dordrecht, The Netherlands: Kluwer Academics Publishers, 2000.
  • 12
    Leder D. Clinical interpretation: the hermeneutics of medicine. Theor Med 1990; 11: 924.
  • 13
    Lock JD. Some aspects of medical hermeneutics: the role of dialectic and narrative. Theor Med 1990; 11: 419.
  • 14
    Bytzer P, Hansen JM, Rune S, et al. Identifying responders to acid suppression in dyspepsia using a random starting day trial. Aliment Pharmacol Ther 2000; 14: 148594.
  • 15
    Mason I, Millar LJ, Sheikh RR, et al. The management of acid related dyspepsia in general practice: a comparison of an omeprazole versus an antacid alginate/ranitidine management strategy. Aliment Pharmacol Ther 1998; 12: 26371.
  • 16
    Goves J, Oldring JK, Kerr D, et al. First line treatment with omeprazole provides an effective and superior alternative strategy in the management of dyspepsia compared to antacid/alginate liquid: a multicentre study in general practice. Aliment Pharmacol Ther 1998; 12: 14757.
  • 17
    Chiba N. Treat the patients’ main dyspepsia complaint, not the Rome criteria (editorial). Am J Gastroenterol 2004; 99: 105962.
    Direct Link: