Consumption of the green flesh kiwifruit (Actinidia deliciosa, Hayward variety) is known to relieve constipation and provide relief from symptoms of digestive dysfunction, but the effect is poorly researched and a credible explanation and mechanism has not been documented. This body of work aimed to increase the understanding of the effects of kiwifruit on the digestion process.

There is some evidence that the mechanism may be due to the proteolytic properties of actinidin, a unique protease in kiwifruit.  Actinidin has a wide pH-dependent reactivity between pH 2.5 - 6 with an optimum at pH 3.0 - 3.5; equivalent to low acid conditions more associated with hypochlorhydria than the pH range of 1.5 - 2.5 normally associated with efficient gastric hydrolysis of protein.  It was hypothesised that in vitro and in vivo the digestion of protein would be facilitated by the presence of the kiwifruit protease, particularly when pH exceeded normal fasting gastric pH. A sequence of studies to measure the effects of lyophilised kiwifruit (KFI), rich in actinidin, on the digestion of protein was undertaken to test this hypothesis.  

Initially, proof of principle of the kiwifruit enzyme activity in vitro was investigated.  Then, the practicality and reliability of a stable isotope technique to measure the rate of gastric protein-digestion in vivo was tested in three studies that compared the effect on gastric emptying dynamics of a protein meal with and without the protease.  In vitro, homogenised freeze-dried meat powder (the protein) was incubated with a series of dilutions of hydrochloric acid (pH 1.3 - 6.4), pepsin (0 - 1.2 % w/w) and KFI (0 - 0.8 % w/w).  Simulated gastric digestion involved incubation at 37ºC for 60 minutes, after which the pH was adjusted to 6.4, and simulated duodenal incubation continued for a further 120 minutes.  The degree of protein hydrolysis was determined using the Kjeldahl technique.  The addition of KFI to the gastric milieu was associated with an increase in the proportion of protein substrate hydrolysed, but only when the concentration of pepsin was sub-optimal for efficient protein hydrolysis and pH of the digestate was elevated above normal fasting gastric pH of >2.5 but <6.4.  Under these conditions the addition of KFI to the milieu doubled protein hydrolysis from a low of 20.7 ± 0.7%, to a high of 55.2% ± 1.2% (mean ± SD).

The aim of the next step described in the thesis was to test the hypothesis that KFI facilitated gastric protein digestion in vivo in a manner similar to the in vitro observations.  As the in vitro study had shown KFI activity was minimal in a simulated duodenal pH > 6.0, this step required a method of measuring the change in the extent of hydrolysis of protein as a result of its passage through the stomach only.  A literature review highlighted many of the problems associated with measuring in vivo factors likely to influence gastric protein-digestion efficiency, including fasting gastric pH, pepsin concentrations and post-prandial gastric re-acidification capacity.  The complexity of measuring any of these factors, combined with limited resources, led to testing the feasibility of measuring an individual's gastric protein-digestion efficiency through an adaptation of the carbon-13 octanoic acid breath test (13C-OABT), as a proxy for protein digestion efficiency.

In the first study, eleven healthy participants aged 58–80 agreed to consume, on two separate occasions one week apart, a three egg white, one yolk omelette with 100 mg of 13C labelled octanoic acid with and without the addition of 2.160 g of KFI (the treatment) taken in capsule form.  On each occasion, two expired breath samples were collected from each participant before they commenced the meal and nine more postprandial over three hours.  The rate of appearance of 13CO2 on the breath, an accepted measure of gastric emptying, is expressed as Tlag (time to maximum rate of detection of isotope on the breath) and T1/2 (time at which 50% of the isotope is retained in the stomach).  The difference in these parameters between the two test meals was hypothesised to reflect the change in gastric protein-digestion efficiency due to the effect of KFI.  Recruitment of participants was not difficult and the procedure was well tolerated.  Based on the findings of the in vitro study, the hypothesis tested was that for participants with normal digestion, treatment would have little or no effect on the rate of protein hydrolysis and would therefore record little or no change in the rate of gastric emptying between control and treatment; whereas protein digestion, for participants with diminished gastric acidification, would be enhanced by the treatment and this would reflect in a one-way change in the dynamics of gastric emptying.  

Of the eleven participants, gastric emptying parameters increased for seven and decreased for four.  The inter-individual variability bought into question the reliability of the test method and the measures were repeated four months later with six participants from the first study.  Variation in repeat measures in the control condition Tlag and T1/2 did not exceed 13% and with the addition of KFI to the meal variation was slightly more at 20–24%.  Again the pattern of some participants exhibiting accelerated gastric emptying and some recording delayed gastric emptying as a result of treatment, was observed.

A third and final in vivo study undertaken in China, employed a non-dispersive infrared spectrometer to measure isotope ratios, as opposed to the isotope ratio mass spectrometer used in the two earlier studies.  It was hypothesised that neither the participant's ethnicity nor the equipment used to measure gastric emptying would significantly alter the measures when compared to studies in New Zealand.  Parameters measured in the twelve volunteers were closely aligned to those of the first study and measures of differences of Tlag and T1/2 with and without treatment were combined (n = 22) into a small meta-analysis.  Treatment with KFI increased Tlag by 7 minutes (95% CI [0, 19], p = 0.060) and T1/2 by 10 minutes (95% CI [0,14], p = 0.027).  However for the 16/22 participants not consuming proton pump inhibitors (PPIs), the effect was more marked: Tlag was increased by 14% or 11 minutes (95% CI [4, 18], p = 0.050) and T1/2 was increased by 10% or 13 minutes (95% CI [4, 23] p = 0.027)  = 0.050) and T1/2 was increased by 10% or 13 minutes (95% CI [4, 23] p = 0.027).  Multiple regression showed that the percentage increase in Tlag with KFI treatment was positively predicted by body weight and negatively predicted by consumption of PPIs. 

Despite trial design shortcomings, the combination of three small in vivo studies, involving 58 individual meals and the collection and analysis of 580 breath samples, indicated KFI had a significant and delaying effect on gastric emptying times (Tlag and T1/2).  This may indicate that for the majority of participants in these studies, KFI improved gastric digestion efficiency, as peptide products of protein hydrolysis are known to delay gastric emptying.  As a result it was hypothesised that the inter-individual variability in response to treatment reflected the underlying gastric protein-digestion efficiency of the individual trial participants, but this hypothesis should be tested in future studies.  

While the novel test method employed in this study requires considerable further research to confirm the validity of the findings, what has emerged from this body of work is evidence that the protease content of the fruit is the likely active ingredient; that the stomach is the likely site of activity and that the effect is likely to be associated with existing gastric protein digestion inefficiency due to sub-optimal peptic hydrolysis.