A Gram Research analysis of 165 TB patients found that a simple blood test measuring nutritional status (PNI score) can predict treatment success with 79% accuracy. Patients with PNI scores below 45 are at high risk for poor treatment outcomes, allowing doctors to provide early nutritional support. The prediction tool showed significant improvement over standard assessment methods, though it requires validation in additional hospitals before widespread clinical use.
Researchers discovered that a simple nutritional blood test called the Prognostic Nutrition Index (PNI) can predict which newly diagnosed tuberculosis patients will struggle with treatment. By studying 165 TB patients over three years, scientists created a prediction tool that identifies high-risk patients early. When PNI scores fall below 45, patients are more likely to have poor treatment outcomes. This finding matters because doctors can now spot at-risk patients quickly and provide extra nutritional support or immune-boosting care before problems develop. The tool showed strong accuracy in testing, though doctors say more research is needed before using it widely in clinics.
Key Statistics
A 2026 research article of 165 newly diagnosed TB patients found that a Prognostic Nutrition Index (PNI) score below 45 predicted poor treatment outcomes, with the prediction tool achieving 78.9% accuracy at 2 months of treatment.
According to research reviewed by Gram, adding the PNI blood test to standard patient assessment significantly improved doctors’ ability to identify at-risk TB patients, with a Net Reclassification Improvement of 0.396 compared to assessment without the test.
A retrospective study of TB patients from 2022-2025 identified optimal PNI cutoff scores of 44.85 at 2 months and 42.6 at 6 months for predicting poor treatment prognosis in newly diagnosed active pulmonary tuberculosis.
The Quick Take
- What they studied: Can a simple blood test measuring nutrition levels predict which TB patients will have trouble responding to treatment?
- Who participated: 165 patients newly diagnosed with active pulmonary tuberculosis between January 2022 and February 2025, all treatment-naive (never treated before)
- Key finding: A PNI score below 45 strongly predicted poor treatment outcomes. The prediction tool achieved 78.9% accuracy at 2 months of treatment, and adding the PNI test significantly improved doctors’ ability to spot at-risk patients.
- What it means for you: If you’re diagnosed with TB, a quick blood test could show whether you’re at higher risk for treatment complications. This allows doctors to give you extra nutritional support early on. However, this tool needs testing in more hospitals before becoming standard practice.
The Research Details
Researchers looked back at medical records from 165 TB patients treated between 2022 and 2025. They measured a blood test called the Prognostic Nutrition Index (PNI), which combines information about protein levels and immune cell counts—basically a score of how well-nourished and immune-ready a patient is. They tracked how well each patient responded to TB treatment at 2 months and 6 months, classifying outcomes as either good (markedly improved) or poor (slightly improved, no improvement, or got worse). Using statistical methods, they found the best PNI cutoff scores that separated patients likely to do well from those likely to struggle. Then they built a prediction tool (called a nomogram) that doctors could use to estimate individual patient risk. They tested this tool’s accuracy using a computer method that simulated 1,000 different versions of their data to make sure results were reliable.
TB is a serious infection that requires months of treatment, and some patients don’t respond well. Catching high-risk patients early means doctors can add nutritional support or immune-boosting treatments alongside standard TB drugs, potentially improving outcomes. A simple blood test is much cheaper and faster than waiting months to see if treatment is working.
This was a retrospective study (looking back at past records), which is less powerful than a prospective study (following patients forward). The sample size of 165 patients is moderate—larger studies would be more convincing. The researchers used proper statistical methods including internal validation with bootstrap resampling, which strengthens confidence in results. However, the study was conducted at one location, so results may not apply everywhere. The authors themselves note that external validation (testing in other hospitals) is needed before widespread use.
What the Results Show
The optimal PNI cutoff score for predicting poor outcomes was 44.85 at 2 months and 42.6 at 6 months of treatment. Patients with PNI scores below these thresholds were significantly more likely to have poor treatment responses. The prediction tool (nomogram) achieved an area under the curve (AUC) of 0.789 at 2 months, meaning it correctly identified at-risk patients about 79% of the time—which is considered good accuracy in medical prediction tools. Adding the PNI test to other patient information significantly improved doctors’ ability to predict outcomes, with a Net Reclassification Improvement (NRI) of 0.396, meaning the test reclassified many patients into more accurate risk categories. The calibration curve showed the tool’s predictions matched actual outcomes well, and decision curve analysis demonstrated the model would provide clinical benefit when used in real practice.
The study found that PNI provided greater value for short-term prognosis (at 2 months) compared to longer-term prediction (at 6 months). This suggests the nutritional status at diagnosis is particularly important for early treatment response. The researchers identified that combining PNI with other patient factors created a more powerful prediction tool than using PNI alone, though PNI was the most important single factor.
Previous research has shown that malnutrition and weak immune function predict poor TB outcomes, but this is the first study to create a specific prediction tool based on the PNI score for newly diagnosed TB patients. The findings align with existing knowledge that nutritional status affects TB treatment success, but provide a practical, measurable way to identify at-risk patients early.
The study looked backward at past medical records rather than following patients forward, which can introduce bias. All patients came from one hospital system, so results may not apply to different populations or healthcare settings. The sample size of 165 is moderate—larger studies would provide more reliable results. The researchers did not test the tool in a completely separate group of patients (external validation), which is the gold standard for proving a prediction tool works. The study did not examine whether early nutritional intervention in high-risk patients actually improved outcomes, only that the prediction tool was accurate.
The Bottom Line
For TB patients: Ask your doctor about a PNI blood test at diagnosis. If your score is below 45, discuss nutritional support options with your healthcare team. For healthcare providers: Consider using PNI testing as part of initial TB assessment to identify high-risk patients who may benefit from enhanced nutritional or immune support. Confidence level: Moderate—the tool shows promise but needs validation in other hospitals before becoming standard practice.
This research is most relevant to: newly diagnosed TB patients (especially those with signs of malnutrition), TB specialists and infectious disease doctors, public health programs in TB-endemic regions, and healthcare systems looking to improve TB treatment outcomes. It’s less immediately relevant to people without TB or those with drug-resistant TB (which may have different risk factors).
The prediction tool works at 2 months into treatment, so doctors could identify at-risk patients and adjust care within the first 2 months. Benefits from nutritional intervention would likely appear over weeks to months as immune function improves, though the study didn’t measure this directly.
Frequently Asked Questions
Can a blood test predict if TB treatment will work for me?
Yes, a nutritional blood test called PNI can predict treatment success with about 79% accuracy. Scores below 45 indicate higher risk of poor outcomes, allowing doctors to provide extra nutritional support early in treatment.
What does a low PNI score mean for TB patients?
A PNI score below 45 means your nutritional status and immune function are lower than ideal, which increases the risk that TB treatment may not work as well. This signals the need for nutritional intervention alongside standard TB medications.
How soon can doctors tell if TB treatment will fail?
The PNI prediction tool works at 2 months into treatment, meaning doctors can identify at-risk patients relatively early and adjust care. This is much faster than waiting 6 months to see if treatment is fully effective.
Is the PNI test ready to use in all hospitals?
Not yet. While the test showed strong accuracy in this study of 165 patients, researchers recommend external validation in other hospitals before widespread use. Ask your TB doctor if they use PNI testing at your facility.
Can improving nutrition help TB treatment work better?
The study shows low nutrition predicts poor outcomes, suggesting nutrition matters for TB treatment success. However, this study didn’t test whether nutritional intervention actually improves outcomes—that requires further research.
Want to Apply This Research?
- Track your PNI score at TB diagnosis and at 2-month and 6-month checkups. Log the specific number (e.g., ‘PNI: 42’) along with your treatment response rating (good vs. poor) to see if the pattern matches predictions.
- If your PNI score is below 45, use the app to set reminders for high-protein meals and nutritional supplements recommended by your doctor. Track daily protein intake and immune-supporting foods (eggs, nuts, beans, fortified grains) to work toward improving your nutritional status during TB treatment.
- Set up monthly check-ins to record your PNI test results and treatment response. Create a simple chart showing your PNI trend over 6 months alongside your TB symptoms and treatment tolerance to identify patterns and share with your doctor.
This research describes a prediction tool for TB treatment outcomes based on nutritional status. It is not a substitute for professional medical diagnosis or treatment. If you have been diagnosed with tuberculosis, work with your healthcare provider to develop an individualized treatment plan. Do not use this information to self-diagnose or delay seeking medical care. The prediction tool described in this study requires external validation before clinical implementation. Always consult your doctor before making changes to TB treatment or starting nutritional supplements, as some may interact with TB medications.
This research translation is published by Gram Research, the science division of Gram, an AI-powered nutrition tracking app.