Can Addition of Doxycycline Perioperatively Reduce Propionibacterium Acnes in Shoulder Arthroplasty?: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Can Addition of Doxycycline Perioperatively Reduce Propionibacterium Acnes in Shoulder Arthroplasty? (NCT03115177), a randomized na study. The primary outcome is Bacterial Culture.

2. Clinical Question

The trial addresses whether adding doxycycline around the time of shoulder arthroplasty is associated with a different frequency of positive intraoperative bacterial cultures than cefazolin alone. The registry description focuses on Propionibacterium acnes, but the primary endpoint description says that cultures were tested for aerobic and anaerobic bacteria and reports participants with a positive bacteria culture. The posted outcome therefore concerns bacterial detection in sampled tissue, rather than a named organism-specific outcome.

The question is clinically situated in the concern that bacteria around the shoulder may be relevant to periprosthetic joint infection, pain, or failure of shoulder arthroplasty. Those concerns motivate the study but are not themselves measured by the posted primary outcome. The record reports culture findings from intraoperative samples and does not report a clinical infection endpoint or a longer-term patient outcome.

Interpretation should stay close to this narrow question: the recorded counts describe positive cultures in the trial groups. The registry record does not report an estimable statistical comparison, confidence interval, or analysis result that would quantify the uncertainty around a difference between regimens.

3. Trial Design

The registry record describes a randomized, parallel-group trial with no masking. It is marked completed, with an enrollment of 56 participants. The recorded study start is 2015-11 and primary completion is 2017-02. The record lists osteoarthritis as the condition and Rush University Medical Center as the sponsor.

Participants were assigned to cefazolin alone or cefazolin with doxycycline. Both groups were described as receiving standard perioperative antibiotics: 2g of cefazolin within 1 hour of incision. The treatment group also received 100mg of intravenous doxycycline within 1 hour of incision. The control group was described as receiving no further antibiotic treatment. These details define the contrast being studied: adding doxycycline to the specified cefazolin regimen, rather than replacing cefazolin with doxycycline.

The trial is described as unmasked. The registry record does not report who knew the assigned treatment, whether outcome assessors were masked, how the random sequence was generated, or how assignments were concealed before enrollment. Those details should not be inferred from the general label of randomized and unmasked.

The primary outcome is titled “Bacterial Culture.” Intraoperative cultures were taken from participants, held for 14 days, and tested for aerobic and anaerobic bacteria. The registry’s timeframe for the outcome is 2 weeks. The posted results list 14 participants with positive cultures in the cefazolin group and 15 in the cefazolin-and-doxycycline group, alongside the respective group sizes of 27 and 29. The record does not provide a statistical analysis estimate for comparing these results.

The phase is listed as NA, and the record reports no secondary outcome measures. It also lists no posted analyses. Details such as an intended sample-size justification, planned interim reviews, and a prespecified statistical analysis plan are not reported in the registry record. The randomized design establishes the intended treatment comparison, but the available result entry does not supply the inferential analysis needed to quantify that comparison.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Bacterial Culture, measured over 2 weeks.

The primary endpoint counts participants whose intraoperative samples had a positive bacterial culture. Cultures were held for 14 days and tested for aerobic and anaerobic bacteria. The outcome timeframe is listed as 2 weeks. This is a participant-level count outcome: it records whether bacteria were detected in cultures, rather than the number of organisms, the amount of bacterial growth, or a clinical diagnosis of infection.

Culture positivity is used because the study’s stated aim is to investigate whether an additional perioperative antibiotic affects bacterial detection around the shoulder. The endpoint is consequently close to the proposed biological target. It can describe whether bacteria were detected in the sampled material under the trial’s culture procedures.

The endpoint does not, by itself, establish whether detected bacteria cause later infection, pain, or implant failure. Nor does a positive result establish that a participant developed a clinical infection. The record does not report an organism-specific breakdown, so the posted outcome cannot be read as a count of participants positive specifically for Propionibacterium acnes. It describes participants with a positive bacteria culture from intraoperative samples.

The results list positive cultures for 14 of 27 participants assigned to cefazolin and 15 of 29 assigned to cefazolin and doxycycline. These are descriptive group results. The registry record does not report an estimable comparison or a confidence interval for the contrast. Without such an estimate, the precision and uncertainty of any difference between the groups cannot be assessed from the posted result entry.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The posted outcome is a count of participants with positive cultures, so interpretation involves both event counts and the number of participants assessed in each group. The registry lists 14 positive cultures among 27 participants in the cefazolin group and 15 among 29 participants in the cefazolin-and-doxycycline group. These are descriptive observations. The record marks a two-arm comparison as not estimable for this count outcome, so it does not provide a trial-based estimate of the contrast.

What does a comparative estimate add?

In general, a comparative estimate puts the outcomes in the randomized groups on a common scale and summarizes how they differ. For a binary outcome such as whether a participant had a positive culture, possible measures include a difference in proportions, a ratio of proportions, or an odds-based measure. Each answers a somewhat different question. The measure intended by the trial’s analysis plan should guide interpretation; the registry record does not report which comparative measure was planned or used.

A raw count is not itself a comparison because group sizes can differ. Group-specific proportions can help describe the observed frequency, but calculating a new contrast from summary values would be a recalculation, not the sponsor’s reported analysis. The registry record identifies the comparison as not estimable, and no recalculated estimate is reported in the registry entries. The descriptive counts should not be converted into a claim about relative effect.

Why does uncertainty matter?

In general, an estimate from a clinical trial is subject to sampling variability: a different set of participants could produce different observed counts, even if the underlying treatment effects were unchanged. A confidence interval communicates this uncertainty around an estimate. A narrow interval indicates greater precision, while a wide interval indicates less precision. A significance test addresses a related but distinct question about compatibility with a specified null hypothesis. Neither an interval nor a test result is reported for this comparison in the registry record.

Failure to report a statistically estimable contrast is not equivalent to a finding of no effect. To support a claim that effects are sufficiently similar, an analysis would need an appropriate estimate and uncertainty assessment, interpreted against a prespecified margin for meaningful difference. The record does not report such a margin or an equivalence or non-inferiority analysis. A non-significant result, even if one had been reported, would not by itself demonstrate equivalence.

What assumptions can matter for a count outcome?

For binary participant-level outcomes, statistical methods generally rely on assumptions about independent observations, the definition and ascertainment of the outcome, and the way participants are analyzed. Randomized assignment is useful for comparing groups, but it does not guarantee that every assumption is met or that missing outcome data are harmless. The record does not explain whether multiple samples per participant were combined into a participant-level result, how culture results were adjudicated, or how incomplete cultures were handled.

Over-dispersion is a general concern when counts vary more than a model expects. It is often discussed for count data that can take many values, such as repeated event counts. The posted primary outcome is instead a count of participants with a positive culture, and the registry record does not report a model or any assessment of over-dispersion. It would therefore be inappropriate to assert that a particular over-dispersion correction was required or used in this trial.

What does the randomized, unmasked design imply?

In general, random assignment aims to balance factors that may affect outcomes, on average, across assigned groups. It supports causal interpretation when assignment is preserved and analysis is appropriate, but it does not prevent chance imbalance or errors in outcome measurement. The trial is listed as randomized and unmasked. The record does not report the randomization procedure, concealment method, or which study personnel were aware of assignment. Because culture is a laboratory outcome, the way samples were collected, labeled, processed, and interpreted can still matter; those procedural details are not reported in the registry record.

Unmasked assignment can influence behavior or measurement in some settings, but the extent of any such influence cannot be determined from the designation alone. The endpoint relies on intraoperative cultures, yet the registry record does not say whether laboratory staff were unaware of treatment allocation. No claim about the presence or absence of bias can be made without those details.

How should planning illustrations be read?

In general, a sample-size illustration based on an observed effect asks how many participants per arm might be needed in a future study to detect an effect under stated assumptions. Such an illustration is not the observed power of the completed trial and does not establish that the trial was adequately powered. The registry record describes a planning approach based on a two-sided significance level of 0.05, equal allocation, a normal approximation, and fractions of the observed effect. It also cautions that these are planning illustrations, not the original trial design. No planning group-size values are reported in the registry record for this analysis, and the observed two-arm effect is not estimable in the registry record. No numerical planning conclusion can therefore be drawn from the posted information.

Most importantly, methods descriptions should distinguish what was planned, what was performed, and what can be reconstructed from summary results. The registry record does not report a statistical analysis plan or an analysis result for the primary comparison. A reader can describe the randomized groups and their culture counts, but cannot attribute an unreported test, adjustment, model, or confidence interval to the trial.

10. Multiplicity, Missing Data and Analysis Populations

The registry record lists one primary outcome measure and no secondary outcome measures. It reports no posted analyses. These record entries describe what is available in the registry; they do not establish whether investigators considered additional outcomes or analyses outside the posted record. The registry record does not report a multiplicity strategy, such as adjustment across multiple hypotheses, and no such strategy should be inferred.

Multiplicity is a general issue when several outcomes, treatment contrasts, subgroups, or interim looks are tested. Considering many analyses can increase the chance of finding an apparently notable result by chance. Prespecification and transparent reporting help readers distinguish the principal question from exploratory analyses. The registry record does not report an analysis plan, interim analyses, subgroup analyses, or a procedure for controlling multiplicity. There is no basis in the posted information for describing how these issues were addressed.

Missing outcome data can affect a group comparison if culture results are unavailable for some participants, particularly when the probability of missingness is related to treatment or outcome status. The listed group sizes and positive culture counts do not explain whether all randomized participants had evaluable cultures, whether any samples were inadequate, or how missing results were handled. The registry record does not report a missing-data method or analysis population definition. It is therefore not possible to confirm from the result entry whether the posted counts represent all randomized participants or another analysis population.

Likewise, the record does not report whether participants were analyzed according to randomized assignment, whether protocol deviations occurred, or whether any per-protocol analysis was conducted. Those details are important for interpreting a randomized comparison, but should not be guessed. The descriptive counts remain the posted result; the registry does not provide enough analysis detail to determine how missingness or population choices affect an estimated treatment contrast.

11. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

The adverse-event entry reports zero serious adverse events in the control group and zero in the treatment group. It also reports zero other adverse events in both groups and zero deaths in both groups. The recorded groups at risk are 27 for control and 29 for treatment. These are the safety figures in the registry record; they are not a comparative estimate with a confidence interval.

A zero count means that no event of that category was recorded in the stated group in this entry. It does not prove that the underlying risk is zero, and it does not establish that the groups have identical safety profiles. With no reported comparative analysis, the precision of a between-group safety difference cannot be assessed. The registry record also does not provide event descriptions, follow-up duration for safety, or information about how events were collected and classified.

Serious and other adverse events are distinct reporting categories. The zero counts should not be interpreted as evidence that no symptoms or minor events occurred beyond what the registry classified and reported. Nor should the absence of reported deaths or serious events be extended to later safety outcomes that are not described in the record. These figures are limited to the posted safety entries and do not support a broader conclusion about safety.

12. What the Result Does, and Does Not, Mean

The result establishes what the registry reports at the level of observed culture counts: 14 participants in the cefazolin group and 15 participants in the cefazolin-and-doxycycline group had positive intraoperative bacterial cultures. The corresponding group sizes are 27 and 29. These figures describe the trial record; they do not provide an estimate of the treatment contrast.

The registry entry states that values are missing or not estimable for a two-arm comparison of this count outcome. Consequently, the result does not establish that adding doxycycline increased or decreased culture positivity. It also does not establish that the regimens were equivalent. To assess the direction, magnitude, and precision of a comparative effect, a reported estimate and its uncertainty would be needed. The registry record does not report those results.

The outcome concerns bacterial detection in intraoperative samples that were held for 14 days and tested for aerobic and anaerobic bacteria. It should not be interpreted as a diagnosis of periprosthetic joint infection or as evidence about pain, implant survival, or arthroplasty failure. The endpoint description does not report an organism-specific count, so it cannot establish the frequency of positive Propionibacterium acnes cultures in particular.

The trial’s stated motivation connects bacteria around the shoulder with potential clinical concerns. That context explains why culture detection was studied, but it does not turn culture positivity into a direct measure of those clinical outcomes. Establishing a relationship between culture results and later patient outcomes would require those outcomes and their analyses to be reported. They are not reported in the registry entry summarized by these registry entries.

The record also lists no secondary outcomes and no posted analyses. Details about missing cultures, analysis populations, and statistical methods are not reported in the registry record. These omissions limit assessment of how the culture counts were analyzed and how precisely any treatment difference could be estimated. The appropriate conclusion is therefore narrow: the trial record contains group-level culture counts, but does not provide an estimable comparative result or evidence about downstream clinical benefit or harm.

13. Why This Trial Matters Statistically

This trial matters statistically because it illustrates the difference between reporting group counts and reporting a treatment comparison. The positive culture counts are available for both randomized groups, but the registry record says that a two-arm comparison is not estimable. A reader should not treat the counts alone as an effect estimate. The denominator in each group and the statistical uncertainty around a contrast both matter.

The endpoint is also a useful example of how an outcome can be close to a biological hypothesis without answering every clinical question. Intraoperative cultures can record bacterial detection under specified sampling and laboratory procedures. They do not, by themselves, measure later infection, symptoms, or implant outcomes. A study can therefore address a mechanistic or microbiological question while leaving patient-centered outcomes unanswered in the posted record.

The randomized design is relevant to the intended comparison, but design labels do not replace analysis details. The record identifies the study as randomized, parallel, and unmasked, while leaving the sequence-generation method, allocation concealment, and analysis population unreported. Readers need both design information and transparent outcome analysis to judge what the result supports.

The safety entries offer another statistical lesson: zero recorded events are observations, not proof of zero underlying risk. Their interpretation depends on the population observed, event ascertainment, follow-up, and uncertainty. Those contextual details are not reported, so the safety counts should remain descriptive.

Finally, the record distinguishes neither an estimable treatment effect nor its precision. A clear statement of that limitation is important: it prevents descriptive data from being mistaken for evidence of equivalence or a treatment effect. The statistical value of the record lies in showing what was counted and making visible what comparative information is not reported.

14. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The primary limitation for interpretation is that the registry record does not report an estimable two-arm comparison for the positive-culture count. The group-level counts are available, but there is no reported effect estimate or confidence interval to show the direction, magnitude, or precision of a contrast. It is not possible to determine from the record whether the issue arose from the analysis, the data available, or another reason.

The outcome is limited to positive bacterial cultures from intraoperative samples. Although the study description identifies Propionibacterium acnes as the motivating organism, the posted primary measure is described as positive bacteria culture and does not include organism-specific results. The record also does not report subsequent clinical infection, pain, implant performance, or arthroplasty failure as outcome results. Culture positivity should not be treated as a substitute for those outcomes.

Important operational details are not reported in the registry record. These include the randomization sequence and concealment, who remained unaware of treatment assignment, culture collection and adjudication procedures, and whether laboratory staff were masked. The record does not explain whether every randomized participant had an evaluable culture or how missing or inadequate samples were handled. Analysis population definitions and protocol deviations are also not reported.

The record reports no secondary outcome measures and no posted analyses. It does not provide a statistical analysis plan, a sample-size justification for the original trial, planned interim analyses, or a multiplicity strategy. The listed planning assumptions relate to illustrations based on observed effects and are not evidence about the trial’s original design. Without these details and an estimable comparison, claims about statistical significance, equivalence, or the strength of evidence are not supported by the posted result.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Can Addition of Doxycycline Perioperatively Reduce Propionibacterium Acnes in Shoulder Arthroplasty? — NCT03115177.
• Registry record last updated 2018-06-15; status last verified 2018-05.
• Results first posted 2018-06-15.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

18. Record Summary

The registry record describes a completed, randomized, parallel, unmasked trial sponsored by Rush University Medical Center. It lists osteoarthritis as the condition, an enrollment of 56, a start of 2015-11, and primary completion in 2017-02. Participants received perioperative cefazolin alone or cefazolin with intravenous doxycycline.

The primary outcome was the number of participants with a positive intraoperative bacterial culture. The record reports 14 positive cultures among 27 participants in the cefazolin group and 15 among 29 participants in the cefazolin-and-doxycycline group. It states that values for a two-arm comparison of this count outcome are missing or not estimable. No secondary outcome measures or analyses are posted. The record does not report a comparative effect estimate, confidence interval, analysis plan, or details on missing culture results. Safety entries list zero serious adverse events, zero other adverse events, and zero deaths in each group.