Praziquantel efficacy: 90% cure in confirmed fluke cases
Praziquantel efficacy is highest when the infection is parasitologically confirmed and the drug is dosed over three consecutive days. In a cohort of 151 people with fluke-related liver lesions, the three-day regimen cured 90.0 percent of confirmed cases. Cure rates climbed slowly, reaching 76.9 percent only at 90 days. Short follow-up windows understate how well the drug works.
Why praziquantel efficacy still leaves practical questions open
Flukes of the genus Clonorchis are foodborne trematodes. People pick them up by eating raw freshwater fish. Chronic infection triggers periductal fibrosis, biliary hyperplasia, and focal eosinophilic infiltration, and the organism sits in the World Health Organization’s Group 1 carcinogen list. On imaging the resulting lesions look like cholangiocarcinoma, so clinicians often start treatment before they can prove what they are treating.
The diagnostic problem is sensitivity. Stool microscopy is the conventional method and it performs poorly in light infections or when the bile duct is obstructed. A patient can present with a textbook exposure history, eosinophilia, and liver lesions, and still return three negative stool samples.
That gap created the question this cohort was built to answer. When you cannot confirm the parasite, does the first-line drug still deliver? Or does the choice of agent matter less than the certainty of the diagnosis?
How the cohort was assembled
The study was a single-center retrospective cohort covering admissions between January 2018 and December 2025. Inclusion required a liver lesion on computed tomography or magnetic resonance imaging, a peripheral eosinophil count above 0.5 x 10^9 per liter, documented raw freshwater fish consumption or residence in an endemic region, and standardized single-agent therapy.
| Design element | Specification |
|---|---|
| Design | Single-center retrospective cohort, 2018 to 2025 |
| Enrolled | 151 patients (78 praziquantel, 73 benzimidazole comparator) |
| Age | Median 34 years (IQR 28 to 42.5) |
| Sex | 84.8 percent male (128 of 151) |
| Entry eosinophil count | Median 7.26 x 10^9 per liter (IQR 3.75 to 13.53) |
| Entry white cell count | Median 15.31 x 10^9 per liter (IQR 10.90 to 20.98) |
| Entry ALT | Median 68 U/L (IQR 42 to 99.5) |
| Lesions per patient | Median 3 (IQR 2 to 3) |
| Praziquantel regimen | 20 mg/kg per dose, three times daily, 3 consecutive days |
| Comparator regimen | 20 mg/kg per day in 2 to 3 divided doses for 7 days |
| Primary outcome | Clinical cure within 90 days |
| Cure definition | Symptom relief plus at least 50 percent lesion shrinkage or resolution, plus eosinophil normalization |
Cure was a composite endpoint and all three components had to be met. Symptoms had to resolve completely or substantially. Follow-up imaging had to show at least 50 percent reduction in the summed lesion diameters, or complete disappearance of the lesions. The eosinophil count had to fall back below 0.5 x 10^9 per liter. Two radiologists read the imaging independently and were blinded to treatment assignment.
Diagnostic certainty split the cohort in two. Definite cases had at least one positive stool examination for the parasite. Suspected cases had three consecutive negative stool examinations alongside typical clinical and radiological features, with no other identifiable cause. The parasitologically definite group came to 30.5 percent of the total, and it was not evenly distributed: 38.5 percent of the praziquantel arm against 21.9 percent of the comparator arm. That difference reached significance at p = 0.027 and had to be handled as confounding.
Overall cure rates
Across the whole cohort the three-day regimen edged ahead. The gap did not clear the significance threshold.
| Outcome | Praziquantel (n = 78) | Comparator (n = 73) | Test |
|---|---|---|---|
| Overall clinical cure | 76.9 percent (60 of 78) | 67.1 percent (49 of 73) | p = 0.179 |
| 95 percent CI | 66.1 to 85.6 | 55.2 to 77.6 | chi-square 1.804 |
| Confirmed cases | 90.0 percent (27 of 30) | 62.5 percent (10 of 16) | p = 0.047 |
| 95 percent CI, confirmed | 73.5 to 97.9 | 35.4 to 84.8 | Fisher exact |
| Suspected cases | 68.8 percent (33 of 48) | 68.4 percent (39 of 57) | p = 0.971 |
| 95 percent CI, suspected | 53.8 to 81.4 | 54.8 to 79.9 | chi-square 0.001 |
| Adjusted odds ratio | Reference | 0.665 | p = 0.318 |
Read the intervals, not the headlines. The comparator’s interval in the confirmed subgroup runs from 35.4 to 84.8 percent, and it is that wide because only 16 people sit behind it. Those 16 people are the reason the authors treat their own conclusion as a hypothesis rather than a finding.
The benefit is concentrated, not spread evenly
Here is the finding that matters most for anyone dosing this drug, and it runs against the intuitive reading.
Split by diagnostic certainty and the picture flips. Where stool examination had confirmed the parasite, the three-day regimen won decisively, 90.0 percent against 62.5 percent. Where stool examination was negative but the clinical picture was typical, the two regimens tied almost exactly, 68.8 percent against 68.4 percent. A multiplicative interaction term for treatment by diagnostic certainty gave an odds ratio of 5.318 with an interval from 0.906 to 31.212 and p = 0.064. That is a trend the authors decline to call a proof.
The mechanism is the useful part. Praziquantel kills adult flukes by disrupting the parasite tegument and triggering calcium influx, and it has minimal activity against juvenile worms. A benzimidazole comparator works differently, binding parasite beta-tubulin and inhibiting both adult and larval stages.
That difference in stage coverage explains the split. A stool-positive patient carries an established adult worm burden, which is exactly what the drug is built for. A stool-negative patient with eosinophilic liver lesions may be carrying juvenile worms, a light burden, or a resolving infection. In that population the stage advantage of an adult-only drug evaporates.
The authors flag a second mechanism. Rapid lysis of adult parasites releases antigen, and that release can aggravate hepatic eosinophilic infiltration before it settles. Onset of benefit and onset of drug action are two different clocks.
Tissue-level resolution is slow but visible. Serial imaging showed one lesion shrinking from 21 by 17 millimeters to 4 by 3 millimeters, and a separate 20 by 17 millimeter lesion disappearing entirely. Histopathology of the lesion areas showed focal necrosis with eosinophil-dominated infiltrate and disrupted lobular architecture. No ova or worms turned up in the sampled tissue.
Timing: the cure curve is the dosing lesson
The longitudinal analysis used generalized estimating equations to account for repeated measures within patients, and it produced the strongest signal in the study. The treatment-by-time interaction was significant at p < 0.001. The two regimens differ in shape, not only in magnitude. After controlling for time, the main effect for the three-day regimen was an odds ratio of 10.278 (95 percent CI 4.123 to 25.62, p < 0.001).
The cumulative cure curve shows why a single early check misleads.
| Timepoint | Praziquantel | Comparator | Note |
|---|---|---|---|
| Day 30 | 7.7 percent (CI 3.0 to 16.0) | 8.2 percent (CI 3.2 to 17.0) | No difference |
| Day 60 | 26.9 percent (CI 17.9 to 38.0) | 31.5 percent (CI 21.5 to 43.3) | Intervals overlap |
| Day 90 | 76.9 percent (CI 66.2 to 85.0) | 67.1 percent (CI 55.4 to 77.0) | Separation emerges |
At one month the two arms are indistinguishable. At two months the comparator is nominally ahead. Only at three months does the intended agent separate. Anyone judging a fluke treatment at day 30 would record a failure in roughly 92 percent of cases. That number is an artifact of the composite endpoint, not a pharmacological result.
One baseline variable predicted cure independently of drug choice: age, with an adjusted odds ratio of 0.615 and p = 0.017. Cure probability falls as age rises. Immunosenescence and established biliary fibrosis are the proposed explanations.
Safety and tolerability
Tolerability was unremarkable. No adverse events were recorded in the three-day regimen arm at all. One patient in the comparator arm had a transient rise in serum creatinine. No serious adverse events occurred in either group.
The authors add a caveat that applies to the safety data specifically. The retrospective design meant adverse events were not systematically documented, so the safety profile is more likely understated than complete. That is a fair reading of any single-center chart review.
Practical recommendations
- Treat the diagnostic tier as a dosing decision and not merely an administrative label. Confirmed infections are where the three-day split-dose regimen earns its place; unconfirmed eosinophilic liver lesions stay a genuinely ambiguous category.
- Assess at 90 days, not 30. A day-30 cure rate near 8 percent tells you nothing about whether the regimen worked.
- Check stage coverage before choosing an agent. Adult-only activity is a limitation whenever juvenile worms may be present.
- Expect antigen-release flare before resolution. Rising eosinophilic infiltrate early in treatment does not automatically mean treatment failure.
- Weight age in prognosis discussions. Older patients have measurably lower cure odds whatever drug is chosen, so age belongs in any praziquantel efficacy forecast.
- Report the interval, not the point estimate, whenever praziquantel efficacy is quoted for a subgroup.
- Do not read a below-threshold lesion result as parasite clearance. Endpoint definitions and assay decision limits are different things.
Limitations
Five constraints bound these results. Treatment was assigned by clinician judgment instead of randomization, so selection bias survives multivariable adjustment. The stool-negative subgroup is not a homogeneous population and may include patients with other eosinophilic liver diseases. Histopathology showed non-specific parasitic eosinophilic inflammation with no adult worms, eggs, or characteristic periductal fibrosis, so histology alone cannot confirm the diagnosis. Adverse events were incompletely documented. The confirmatory subgroup held only 46 patients, split 30 and 16 between the arms, which is what makes those confidence intervals so wide.
One reporting detail deserves flagging. The narrative text reports individual lesion dimensions in millimeters, while a baseline characteristics table lists lesion size without a consistent unit. The millimeter figures from the narrative are used here because they line up with the imaging follow-up.
Bottom line
Praziquantel efficacy in this cohort is real but conditional. The three-day, 20 mg/kg three-times-daily regimen produced a 90.0 percent cure rate in parasitologically confirmed fluke infections and no adverse events, while delivering no measurable advantage over a comparator in stool-negative suspected cases. The drug’s limitation is stage coverage rather than potency, and that distinction is what makes praziquantel efficacy so dependent on when it is measured. Dose the confirmed case confidently. Assess at 90 days instead of 30. Treat the unconfirmed eosinophilic lesion as an open diagnostic question rather than a settled treatment target.
FAQ
What is the standard praziquantel dose for liver fluke infection?
The regimen used in this cohort was 20 mg/kg per dose, given three times daily for 3 consecutive days, for a cumulative exposure of 180 mg/kg. That split-dose structure is standard because it maintains drug levels long enough to cover adult flukes in the biliary tree. A single large dose is not equivalent, and a shorter course leaves adult worms in place.
How effective is praziquantel for confirmed fluke infection?
In parasitologically confirmed cases the cure rate was 90.0 percent, or 27 of 30 patients, with a confidence interval of 73.5 to 97.9 percent. Cure required symptom resolution, at least 50 percent lesion shrinkage on imaging, and eosinophil normalization below 0.5 x 10^9 per liter, all within 90 days. That is the highest praziquantel efficacy figure in the cohort, and it comes from the most demanding endpoint definition.
Why does praziquantel not work as well in suspected cases?
Because the drug acts on adult flukes and has minimal activity against juvenile worms. A stool-negative patient may be carrying larvae, a light burden, or a resolving infection rather than established adults. In that population the stage advantage disappears, which is why cure rates in suspected cases were nearly identical between the two regimens studied.
How long does praziquantel take to show an effect?
Slower than most clinicians expect. Cumulative cure was 7.7 percent at day 30, 26.9 percent at day 60, and 76.9 percent at day 90. The lag reflects how cure was defined, since lesion shrinkage and eosinophil normalization both take weeks to occur. Judging the drug at one month will make an effective regimen look like a failure.
Can praziquantel make symptoms worse before they improve?
Yes, transiently. Rapid lysis of adult flukes releases antigen, and that release can aggravate hepatic eosinophilic infiltration before the lesion resolves. Worsening inflammatory markers early in treatment are therefore not automatically a sign that the drug failed. They are a predictable consequence of killing a large parasite burden quickly.
What are the side effects of praziquantel at this dose?
In this cohort, none were recorded in the three-day regimen arm. One patient in the comparator arm had a transient creatinine elevation, and no serious adverse events occurred in either group. The authors caution that retrospective chart review under-documents adverse events, so absence of reported events is weaker evidence than prospective monitoring would be.
Who responds less well to praziquantel?
Older patients. Age was the only baseline variable that independently predicted cure, with an adjusted odds ratio of 0.615 and p = 0.017, meaning cure odds fall as age rises. Age-related immune decline and established biliary fibrosis are the proposed mechanisms. Age did not interact with treatment choice, so it affects prognosis under either regimen.
How should cure be assessed after praziquantel treatment?
At 90 days, using imaging plus laboratory and clinical criteria together. Cure in this study required all three: symptom relief, at least 50 percent reduction in summed lesion diameters or complete resolution, and eosinophil count return below 0.5 x 10^9 per liter. Checking only one domain, or checking too early, produces a misleading answer.
Why is stool testing unreliable for these infections?
Sensitivity is low in light infections and in patients with biliary obstruction, so false negatives are frequent. The reference standard used here required at least one positive stool examination, and suspected cases required three consecutive negatives. Even so, a negative result does not exclude infection, which is precisely why empirical treatment decisions get made.
Does lesion size predict whether treatment will work?
Not in this cohort. Maximum lesion diameter was not an independent predictor of cure in the generalized estimating equations model, and neither was lesion number. What did predict outcome was diagnostic certainty and patient age. Lesion size remains useful for tracking response over time, but it does not forecast it.
How quickly are lesions resolved on imaging?
Individual lesions shrank substantially over the observation period. One lesion measured 21 by 17 millimeters at baseline and 4 by 3 millimeters after treatment, and a separate 20 by 17 millimeter lesion resolved entirely. The cohort as a whole showed a gradual rather than abrupt response, consistent with the 90-day composite endpoint used to define cure.
Can praziquantel be combined with other antiparasitic drugs?
Not tested here. Every patient in this cohort received single-agent therapy, and combined use of other anthelmintics or corticosteroids during treatment was an explicit exclusion criterion. The authors call for prospective trials of combination and sequential regimens, which means the sequential question remains unanswered rather than answered negatively.
