When a COPD patient needs oxygen: GOLD guidelines applied to Indian practice

Long-term oxygen therapy (LTOT) is one of only three interventions — alongside smoking cessation and, in selected patients, lung volume reduction — that has been shown to reduce mortality in COPD. The evidence is four decades old and has not been overturned. Yet in Indian practice, LTOT is prescribed inconsistently: some patients who would benefit are never offered it, some patients who are started on it use it for three or four hours a day because nobody explained that dose matters, and some patients who do not meet criteria are on supplemental oxygen anyway because a well-meaning physician did not want to refuse a breathless family. This article sets out the GOLD 2024 criteria clearly, explains the evidence base, and addresses the Indian reality of prescribing LTOT in a system where arterial blood gases are not universally available.

The target reader is the physician or respiratory therapist actively making LTOT decisions, the pharmacy-chain staff member receiving prescriptions that may or may not be complete, and the patient or family member trying to understand why a pulmonologist did or did not recommend oxygen.

The evidence base: NOTT and MRC

Two trials, both completed in 1980–1981, define modern LTOT.

The Nocturnal Oxygen Therapy Trial (NOTT) enrolled 203 patients with severe COPD and documented arterial hypoxaemia (PaO₂ ≤55 mmHg, or 55–59 mmHg with evidence of cor pulmonale or polycythaemia). Patients were randomised to nocturnal oxygen only (~12 hours/night) or continuous oxygen (averaging ~18 hours/day). The continuous-oxygen group had approximately half the mortality of the nocturnal-only group at 24 months. The dose-response within the study — more hours of oxygen, lower mortality — was the critical finding.

The Medical Research Council (MRC) trial enrolled 87 patients with severe COPD and chronic hypoxaemia. Patients were randomised to 15 hours/day of oxygen (including overnight) or to no supplemental oxygen. At 5 years, the treatment group had ~45% mortality versus ~67% in controls.

The combined interpretation that has held since 1981: in COPD patients with documented arterial hypoxaemia, supplemental oxygen reduces mortality, the effect is dose-dependent, and the minimum dose that produces clinical benefit is approximately 15 hours per day. Fewer than 15 hours per day does not appear to produce a mortality benefit. More than 15 hours produces a larger benefit, with the NOTT comparison suggesting ~18 hours per day is better than 15.

No subsequent trial has overturned these findings. The LOTT trial (2016) tested supplemental oxygen in patients with moderate hypoxaemia (SpO₂ 89–93%) and did not show a mortality or hospitalisation benefit. The result reinforced, rather than challenged, the NOTT/MRC thresholds: LTOT helps severe hypoxaemia; it does not help moderate hypoxaemia.

GOLD 2024 criteria

The Global Initiative for Chronic Obstructive Lung Disease (GOLD) 2024 report restates the LTOT indication as follows (GOLD Report):

Prescribe LTOT for stable COPD patients with:

1. PaO₂ ≤55 mmHg (SaO₂ ≤88%) at rest, breathing room air, measured at least twice, at least three weeks apart, during a period of clinical stability;

or

2. PaO₂ 55–60 mmHg (SaO₂ ~89%) with at least one of:
   • Evidence of cor pulmonale (ECG, echocardiographic, or clinical)
   • Polycythaemia (haematocrit >55% or haemoglobin >17 g/dL)
   • Pulmonary hypertension
   • Peripheral oedema suggesting right heart failure

The patient must be stable — not within six weeks of an exacerbation, on optimised medical therapy (long-acting bronchodilators, inhaled corticosteroids if indicated), and not smoking. The target of therapy is SaO₂ ≥90% (SpO₂ ~92%) during rest, sleep, and exertion.

The minimum duration for mortality benefit is 15 hours per day; GOLD and ATS/ERS recommend 15 or more hours per day, with most guidelines pointing at 15–24 hours. Practically, patients are told to wear oxygen “while sleeping, while awake at home, during any activity that brings on breathlessness, and during meals” — which, accumulated across the day, tends to hit 15–18 hours.

GOLD 2024 explicitly distinguishes LTOT (chronic, continuous therapy for severe resting hypoxaemia) from:

• Nocturnal oxygen only — for patients who desaturate primarily during sleep without meeting LTOT criteria awake. Evidence for mortality benefit in this group is weaker.
• Ambulatory oxygen — for patients who desaturate on exertion but not at rest. GOLD does not recommend routine ambulatory oxygen for non-hypoxaemic patients based on the LOTT findings, but it remains appropriate in some subgroups.
• Short-burst oxygen (e.g. pre- or post-exercise, palliative). GOLD notes this is widely prescribed but poorly evidenced.

The clinical position is unambiguous: LTOT in COPD is for arterial hypoxaemia. PaO₂ 55 mmHg or SaO₂ 88% is not a gray zone. Below that, therapy; above that, therapy only with the specific comorbidities listed.

Applying the criteria in Indian practice

The GOLD criteria assume access to stable-state ABG measurements. Indian practice operates with variable ABG access, and the working adaptation is broadly as follows.

The ABG-available scenario

At tertiary centres (AIIMS network, major private hospitals, tier-1 city tertiary care), a patient with stable GOLD 3–4 COPD is referred for an ABG during an outpatient visit. Two ABGs, three weeks apart, both showing PaO₂ ≤55 mmHg, establish LTOT indication. The prescription specifies flow (typically 1–3 LPM), duration (≥15 hours/day, often written as “continuous except while bathing”), target SpO₂ (≥90%), and device (oxygen concentrator for home). Follow-up is at 6–12 weeks with a repeat ABG on oxygen to confirm adequate correction without CO₂ retention.

The ABG-unavailable scenario

In much of India — smaller hospitals, rural primary care, most tier-2 and tier-3 private clinics — ABG is not available at the point of prescribing. The working compromise is SpO₂-based triage:

• SpO₂ <88% at rest, room air, in a stable patient, on two occasions three weeks apart is taken as a surrogate for PaO₂ ≤55 mmHg.
• The correspondence is imperfect. SpO₂ 88% can correspond to PaO₂ ranging from ~52 to ~60 mmHg depending on pH, temperature, CO₂, and curve-shift factors. It is also subject to the pulse-oximeter accuracy issues covered in understanding SpO₂ vs PaO₂ vs SaO₂ — particularly the dark-skin over-reading effect.
• In borderline cases (SpO₂ 87–90%), multiple readings across different fingers, perfusion states, and oximeters give a more defensible picture.
• When any clinical finding suggests CO₂ retention — morning headaches, daytime somnolence, plethora, cor pulmonale — the patient should be referred for ABG before or during LTOT initiation, even if that means a trip to a district or referral hospital.

For the prescriber, the SpO₂-only LTOT initiation is provisional. It is ethically defensible in a system where ABG access is a barrier, but the prescription should be reviewed — and, ideally, confirmed with an ABG — within 90 days.

What Indian GOLD adaptation does not say

There is no universally adopted Indian altitude modifier for LTOT criteria. A resident of Leh or Manali has lower baseline PaO₂ than a Mumbai resident; the GOLD thresholds are sea-level thresholds. Practitioners at altitude generally apply sea-level thresholds with clinical judgement — a Leh resident with PaO₂ 50 mmHg on ABG would likely benefit from LTOT, but the discussion of risk-benefit shifts because the baseline is lower to begin with.

The Indian Chest Society’s LTOT guidelines are broadly aligned with GOLD, with some additional emphasis on:

• Ensuring tuberculosis has been ruled out or adequately treated before attributing hypoxaemia to COPD (Indian context — TB remains a major confounder in chronic respiratory disease).
• Smoking cessation as a hard precondition for LTOT reimbursement.
• Monthly clinical follow-up during the first three months of LTOT.

COPD prevalence and the care-delivery gap

The Indian Study on Epidemiology of Asthma, Respiratory Symptoms and Chronic Bronchitis (INSEARCH) estimated COPD prevalence at approximately 7.4% in adults aged 35 and above. The Global Burden of Disease estimates for India place the adult COPD prevalence at 4–8% depending on definition, with substantial variation across states — Himalayan and Northeast states tend to have higher prevalence because of indoor biomass burning in addition to tobacco. The absolute numbers are large: in a population of over 1.4 billion, even a 5% adult prevalence is in the tens of millions.

Of these, the fraction with severe GOLD 3–4 disease and qualifying arterial hypoxaemia is in the low single-digit percent of the COPD population — plausibly 1–3 million patients eligible for LTOT. Actual LTOT penetration in India is far below this number. Oxygen-concentrator sales data and pharmacy-dispensing patterns suggest that perhaps 100,000–300,000 patients are on home LTOT at any given time — an order of magnitude below eligibility.

The gap reflects:

• Under-diagnosis of severe COPD. Spirometry is not routinely available in primary care; many patients are managed for “chronic bronchitis” without a spirometric COPD diagnosis.
• Under-referral for ABG. Where ABG would establish the indication, the referral does not happen.
• Cost. An oxygen concentrator retails at ₹35,000–₹1,20,000 in the Indian market; ongoing electricity costs for 15+ hours daily use add ₹400–₹1,200 per month. For a patient with monthly household income below ₹20,000, this is prohibitive without reimbursement.
• Reimbursement patchiness. CGHS generally reimburses oxygen concentrators for qualifying beneficiaries with documented indications; ECHS varies by hospital; private insurance typically does not cover durable medical equipment; PMJAY does not cover home oxygen therapy as a dedicated benefit. The patient most likely to have LTOT is a government or armed-forces retiree with CGHS/ECHS; the rural PMJAY beneficiary is not covered.
• The compliance ceiling. Once prescribed, many patients use oxygen for far fewer than 15 hours per day. A prescription of “continuous” interpreted by the patient as “when I feel breathless” does not produce mortality benefit.

Practical prescribing decisions

The prescriber writing an LTOT prescription should specify, explicitly:

1. Flow rate (LPM). For most COPD-LTOT patients, 1–3 LPM is adequate. Start at 1 LPM if the resting PaO₂ is 55–60 mmHg; start at 2 LPM if PaO₂ is <55 mmHg. Titrate to SpO₂ 90–93%.
2. Duration (hours/day). The number should be ≥15 and ideally 18–24. Write it explicitly — “at least 16 hours/day including the entire sleep period” is clearer than “continuous”.
3. Device type. A stationary concentrator is the workhorse. See the oxygen concentrator reviews on this site for specifications and long-term reliability reports. A small cylinder or portable concentrator may be added for mobility outside the home; this is an adjunct, not a primary device, because of cost and sustainability.
4. Target SpO₂ and when to adjust. SpO₂ 90–93% during rest on oxygen. Above 93% at a given flow means flow can be reduced; below 90% means flow should be increased or the patient re-evaluated.
5. CO₂ retention check. Patients with hypercapnic COPD need monitoring. If the patient has any feature suggesting CO₂ retention (morning headache, daytime somnolence, confusion on initiation), an ABG on oxygen is indicated to confirm that oxygen is not worsening CO₂ without also reducing the respiratory drive.
6. Follow-up. Clinical review at 4–8 weeks, repeat ABG at 8–12 weeks, and an annual review thereafter.

A prescription that does not contain flow, duration, and target is incomplete. Patients and families routinely report prescriptions that read “oxygen as required” — this is a ticket to undertreatment.

Why LTOT compliance is load-bearing

The NOTT finding that ~18 hours per day outperforms ~12 hours per day is not a gentle slope. It is the core of LTOT — the intervention is effective at the dose it was tested at, and progressively less effective below that dose. Three clinical habits predict compliance success:

• Machine accessible. A concentrator that lives in a separate room with a short tube and no extension is used less than one with an appropriate extension tube (12–15 m) that lets the patient move around the house without disconnecting.
• Patient understands the dose. A patient who is told “use it when you feel short of breath” will average 4–6 hours/day. A patient who is told “this will extend your life by two years at 18 hours/day and will not help you at all at 4 hours/day” often hits 15 hours. Honesty with numbers is not cruelty.
• Family buys in. Home LTOT sustains on family compliance. In joint households, the primary caregiver’s understanding of why the machine needs to be on during sleep and meals is the single biggest predictor of real-world duration.

Consult your pulmonologist before discontinuing or reducing prescribed LTOT hours — once started, the therapy is cumulative in its mortality effect, and gaps in delivery reduce its value in ways the patient cannot always feel.

Closing: the threshold is not negotiable

PaO₂ 55 mmHg (SaO₂ 88%) is not a clinical opinion. It is the threshold below which randomised-trial evidence establishes mortality benefit from supplemental oxygen, and above which (in patients without cor pulmonale or polycythaemia) it does not. A patient whose PaO₂ is 62 mmHg does not need LTOT, no matter how breathless they feel — breathlessness at that PaO₂ is typically correctable with optimised bronchodilator and inhaled corticosteroid therapy and pulmonary rehabilitation, not oxygen. A patient whose PaO₂ is 54 mmHg needs LTOT, even if they do not feel catastrophically unwell. The number is the thing.

Indian practice sits under an evidence base it can largely meet — the main barriers are diagnostic access, device cost, and compliance after prescription. Each is addressable. The failure mode is not a prescription technicality; it is a patient who qualifies for LTOT, does not receive it, and dies younger than they should.

Primary references that inform clinical practice in this area: GOLD 2024 Report; Indian Chest Society Guidelines for LTOT 2017; NOTT 1980; MRC 1981; LOTT 2016; INSEARCH 2012.