Dietary iron does not impact the quality of life of patients with quiescent ulcerative colitis: an observational study

Anaemia due to iron deficiency is prevalent amongst patients with inflammatory bowel disease (IBD) [1‐3]. Although prescription of oral iron therapy is common practice it may, on occasions, undermine the intended benefits of therapy by simultaneously aggravating symptoms and thereby possibly reducing quality of life in these patients [4]. For patients who are not in stable remission, aggravation of gastrointestinal symptoms could occur through the overproduction of harmful reactive oxygen species in the colon, catalysed by a surplus of unabsorbed supplemental iron [5]. Indeed, in murine models of IBD, high levels of dietary iron increase inflammation and colonic lipid peroxidation, and induce mucosal damage [6, 7]. Furthermore, IBD patients have an increased risk of developing colorectal cancer [8] and, worryingly, recent murine studies have suggested strong promotion of intestinal tumorigenesis with an increase in luminal iron concentration [9, 10].

Reactive (or ‘free’) iron in the intestinal lumen has not been characterised. However, even for supplementation with soluble ferrous (Fe(II)) iron salts most of the iron will not remain in the lumen in a readily reactive form. Oxidation, and then oxo-hydroxide precipitation, in the gut lumen will greatly limit the ‘free’ iron. Thus free reactive iron in the gut lumen, that can exacerbate inflammation, may occur at only a low concentration. In light of this, the potential adverse effects of iron at normal (or near normal) fortified dietary iron intake levels have been investigated albeit data are still scarce. Nonetheless, all suggest a potential detrimental effect of iron. An intervention study in healthy, non-anaemic middle-age men using flour fortified with 30 mg Fe/kg as ferrous sulphate for more than 1 year has shown a significant increase in systemic biomarkers of oxidative stress such as superoxide dismutase and glutathione peroxidase [11]. Zimmermann et al. have reported significant changes towards a more detrimental gut microflora profile in children supplemented with electrolytic iron-fortified biscuits for 6 months [12]. To our knowledge, the only human study in IBD is our own recently published data indicating a detrimental effect of dietary iron, in particular from Fe-fortified products, on the quality of life of IBD patients with mild-moderately active disease [13]. The aim of the observational study presented herein was to determine the effect of dietary iron intakes on the quality of life of ulcerative colitis (UC) patients with inactive (i.e. quiescent) disease for at least 1 month. In particular, and unlike in our previous work, we used a single, defined and larger (n = 50) IBD population (quiescent UC) and used the most robust, albeit resource-consuming, dietary assessment technique, namely 7-day food diaries. As a reference group we also studied healthy controls and they, for analysis, were matched to subjects with UC.

A total of 103 subjects were enrolled in the study (Figure 1). Of the UC cases that completed the study 26 were male (54%) and 22 were female (46%) with a mean age (range) of 50 (23–78) years. Of the healthy controls 26 were male (49%) and 27 were female (51%) with a mean age (range) of 41 (20–76) years. The differences between groups were not statistically significant. Six UC patients relapsed during the study and were, therefore, excluded from the main analysis. Of the remaining UC patients 34 were matched for age (within 5 years) and gender with controls.

The mean EuroQol VAS score for the UC patients with quiescent disease (i.e. in remission) was 81% (95% CI = 76–85), which was significantly lower than that of matched controls (87%: 95% CI = 84–91) (p = 0.007). Dietary intake of all ‘forms’ of iron (i.e. total, haem, non-haem and fortificant) had no statistical significant linear association with the EuroQol VAS score in the matched cases and controls (Figure 2). Similarly when the unmatched datasets were compared in their entirety (i.e. 42 UC patients in remission and 53 controls) there was again no difference between the two groups and no association with dietary iron intakes. The same was true when using the EuroQol index value (Additional file 1: Figure S1) or the IBDQ score (patients only) as quality of life measures (Figure 3). Multiple linear regression, with fortificant iron, haem iron and ‘natural’ (i.e. non-fortificant, dietary-derived) non-haem iron as independent variables (predictors) and the quality of life measure (i.e. EQ VAS, EQ index or IBDQ score) as the dependent variable, also showed no association. Furthermore, there were no statistically significant differences for the intake of all ‘forms’ of dietary iron between cases and controls in the matched study group. A summary of the dietary intake of the main nutrients in the matched cases and controls is presented in Additional file 2: Table S1.

Although not part of the above main analysis we briefly considered the six patients who relapsed during the study period. Notably, these patients had significantly lower dietary intakes of total and non-haem iron compared to patients in remission (i.e. the cases in the matched study group), even though total daily energy intakes and general diet composition were comparable (Table 1). Unsurprisingly, the relapsing patients also had lower quality of life scores (Table 1) and, splitting the analysis into the four IBDQ domains (i.e. bowel, systemic, emotional and social) showed that, while all domains were associated with relapse of disease, bowel and emotional function scores dropped the most (Table 2). As observed previously [13] for the 6 patients who relapsed there was an inverse trend towards lower IBDQ scores (i.e. quality of life) with higher total dietary iron intakes (p = 0.029 for the comparison of the slopes of the linear associations for active and inactive disease, Figure 3). The significance of this difference was further strengthened when the IBDQ scores were adjusted for disease activity and total dietary iron was expressed corrected for daily energy intakes (p = 0.0001, Figure 4). Numbers were too small to consider subdividing the analysis for the different forms of dietary iron.

Our previous data have shown that elevated dietary iron intakes, especially fortificant iron intakes, are associated with a reduced quality of life in IBD patients with mild-moderate disease activity [13]. This implies that during relapse, luminal iron levels become an attributable factor in the severity of inflammation. This is supported by supplementation data in rodent models over the past 10 years as well as more recent work showing that depletion of luminal iron alleviates gut inflammation in IBD models [6, 7, 28].

The aim of the present study was to investigate if, in the absence of inflammation (i.e. in inactive (quiescent) disease), this might still be the case. Our results show that, in the absence of clinical disease activity (i.e. SCCAI < 5), there was no measurable linear association between dietary iron intakes and quality of life which is consistent with earlier preliminary findings [13].

However, even though the numbers are very small (n = 6), our data in UC patients that relapsed after enrolling in the study, and developed mild-moderate active disease, again showed a negative association between dietary iron intake and the disease-specific quality of life score (IBDQ). This very small number of patients would not justify further sub-analysis but we do note that, for all UC patients studied herein, the IBDQ scores gave a good correlation with EuroQol scores, especially EQ VAS(%) (Additional file 3: Figure S2). The small dataset also showed that mean dietary iron intakes were lower in the UC patients with mild-moderate active disease compared with their counterparts in remission and this is in agreement with previous studies in IBD [29]. It suggests that the disease state, itself, modifies patient food choice.

The use of 7-day food diet diaries and carefully matched control data means that we can suggest confidently, from this work that dietary iron does not have any clinically meaningful effect on quality of life in UC patients with inactive disease. However, the picture appears different in the presence of existing mucosal inflammation: some degree of disease activity, even if mild (SCCAI ≥ 5), may be sufficient to allow a detectable association between quality of life and dietary iron intake (Figure 3A and [13]). Moreover, this appears to be driven by fortificant iron [13]. Taken together, there may also be important resonances between our studies (here and [13]) and the recent work of Radulescu et al.[9] where dietary iron was shown to be a permissive factor for colon cancer development in murine models.

We propose that luminal free iron may be a contributing factor in the severity of colonic inflammation, once triggered through another means, in patients with UC. We do acknowledge that data from the robust observational study presented here only provide suggestion of associations and a more conclusive intervention study in patients with active disease is now required. If the association shown here and in our prior preliminary work [13] is confirmed, then, whether it operates through the well described reactive oxygen species-inducing nature of reactive iron in the colon, or through iron’s modulation of the microbiota [7], or indeed some other mechanisms(s) would warrant careful investigation, as would the possibility of intervention through sequestration of free luminal iron in the distal bowel.