Cold chain (revision 50)

A cold chain is an unbroken sequence of temperature-controlled storage, handling and transport operations that maintains a temperature-sensitive product within its labelled storage conditions from the point of manufacture to the point of use. The term is used both for the physical arrangement — refrigerated rooms, insulated shippers, coolant, vehicles — and for the documentary system of records and custody transfers that demonstrates the arrangement worked.^[1]

The concept originates in vaccine distribution, where it was developed from the 1970s onward as part of expanded immunisation programmes, and it retains much of that vocabulary. It has since been generalised to any product whose labelled storage conditions are narrower than ambient, including insulin and the GLP-1 receptor agonists, most of which are labelled for storage at 2–8 °C before first use.^[2]

Two properties distinguish a cold chain from ordinary logistics. It is a chain in the strict sense that its integrity is set by its weakest link rather than by an average: an hour on an unshaded loading dock is not offset by a week of correct refrigeration. And its principal failure is invisible on inspection, because temperature-mediated degradation of a peptide or protein product generally produces no visible change. This combination is why monitoring instrumentation, rather than examination of the product, carries the evidentiary burden.^[3]

The commonest documented failure of pharmaceutical cold chains is not overheating but inadvertent freezing. Systematic reviews of vaccine distribution have repeatedly found freezing exposure in the majority of studies examining shipments and storage, in high-income and low-income settings alike, and have identified it as more prevalent than heat exposure. Freezing is more damaging than mild warming for many liquid biological products and is not detectable after the fact by inspection.^[4][5]

Labelled storage statements draw on compendial definitions rather than on plain language, and the definitions are narrower than everyday usage suggests. USP General Chapter <659> defines the terms used on United States labelling, and the European Pharmacopoeia and WHO guidance use closely comparable ranges.^[6][1]

Two features of this table are consequential and routinely missed. Controlled room temperature is not a synonym for whatever temperature a room happens to be; it is a specification with a permitted excursion band and a mean kinetic temperature ceiling, and a warehouse that averages 27 °C does not satisfy it. And the definitions of controlled cold and controlled room temperature both build permitted excursions into the definition itself, so a brief departure from the nominal band is not automatically an excursion in the regulatory sense — a distinction developed at Temperature excursion.^[6]

Storage statements for the compounds covered on this wiki cluster in two groups. Manufactured injectable GLP-1 receptor agonists are labelled for cold storage before first use, with an in-use period at higher temperature after first use. Lyophilised research peptides are commonly accompanied by a recommendation of frozen or cold storage for long-term holding and a statement that the dry material tolerates ambient transit, a combination whose physical basis is set out at Lyophilisation and whose evidential basis is generally absent.^[7]

The World Health Organization additionally defines a controlled temperature chain for specified vaccines, permitting a single excursion to ambient temperatures up to 40 °C for a defined period immediately before administration, subject to product-specific stability data and to a monitoring device that records the exposure. It is a deliberate, evidence-supported relaxation of the 2–8 °C requirement rather than a tolerance of failure, and it illustrates that cold-chain requirements are properties of products rather than of logistics.^[3]

A cold chain produces a record, and the instrument that produces it determines what the record can support.

Electronic data loggers are the reference instrument for shipments and for storage. A logger samples a thermistor or thermocouple at a fixed interval, stores time-stamped values in non-volatile memory, and reports either on retrieval or, in connected variants, in near real time. What matters for interpretation is the sampling interval, the accuracy and its temperature dependence, the calibration status, and whether the sensor is in the payload or in the air of the container. A logger taped to the outside of a shipper records the transit environment; a logger buried in the payload records something much closer to what the product experienced, and the two can differ by several degrees for hours.^[8]

Simpler devices remain in use and answer narrower questions:

Minimum–maximum thermometers

Record the extremes reached but not their duration, so they cannot support any cumulative calculation.

Chemical threshold indicators

Change colour irreversibly when a threshold is crossed. Freeze indicators of this type are the standard means of detecting a freezing excursion in vaccine distribution.

Time–temperature integrating indicators

Accumulate a colour change as a function of both temperature and time, approximating the cumulative thermal exposure that matters for degradation.

The vaccine vial monitor deserves separate mention because it is attached to the individual container rather than to the shipment, and therefore survives every custody transfer. It is a heat-sensitive label whose inner square darkens progressively; when it matches or exceeds the reference ring the vial is not to be used. Its limitation is definitional: it responds to heat and not to freezing, so a vial that has been frozen presents an unchanged monitor.^[3]

A temperature record is only as good as the calibration of the instrument that produced it. Compendial guidance on monitoring devices expects calibration traceable to a national metrology institute, at intervals appropriate to the device, and at points within the range of interest rather than only at ambient temperature. A logger calibrated at 25 °C and used at 5 °C may carry an error larger than its stated accuracy.^[8]

The conventional expectation for a logger used in 2–8 °C distribution is an accuracy of ±0.5 °C across that range. This is an expectation drawn from qualification practice and from performance specifications for prequalified devices; it is not a requirement imposed by a general chapter, and articles asserting it as a standard have been corrected on this point.^[1][8]

Traceability of the record itself is a separate question from calibration of the sensor. A record that cannot be tied to a specific consignment, with the times of custody transfer identified, cannot be used to attribute an excursion to a stage of the chain, which is the purpose for which such records are chiefly wanted.^[9]

Cumulative thermal exposure cannot be summarised by an arithmetic mean, because degradation rate rises exponentially with temperature. Time spent warm contributes disproportionately, and an average conceals it. The conventional summary statistic is the mean kinetic temperature, introduced by Haynes in 1971: the single constant temperature that would produce the same extent of degradation, over the same period, as the varying profile actually experienced.^[10]

For a series of intervals of duration t_i at absolute temperatures T_i, the mean kinetic temperature is

T_MKT = (ΔH/R) ÷ [ −ln( Σ t_i e^{−ΔH/(R T_i)} ÷ Σ t_i ) ]

where ΔH is an assumed activation energy and R the gas constant. Compendial practice takes ΔH as 83.144 kJ·mol⁻¹, chosen so that ΔH/R is exactly 10,000 K, which simplifies hand calculation and is the value used throughout this wiki.^[11][10]

A worked case shows why the statistic is used. Consider a warehouse that spends 12 hours of each day at 2 °C and 12 hours at 30 °C — an unrealistically extreme cycle, chosen to make the arithmetic legible. The arithmetic mean is 16.0 °C. Converting to kelvin, 275.15 K and 303.15 K:

e^{−10000/275.15} = e^−36.3438 and e^{−10000/303.15} = e^−32.9869

The first term is smaller than the second by a factor of e^−3.357, or 0.0349, so it contributes almost nothing. Their mean is e^−32.9869 × 0.5174, and

T_MKT = 10000 ÷ (32.9869 + 0.6590) = 10000 ÷ 33.6459 = 297.21 K = 24.1 °C

The mean kinetic temperature is 24.1 °C, some 8 °C above the arithmetic mean and only 6 °C below the maximum. Half the time at 2 °C bought almost nothing, because at that temperature almost no degradation occurred and the statistic weights by degradation, not by clock time.^[10]

Three cautions attach to the statistic and are stated in the compendial guidance itself. It presumes a single Arrhenius-behaved degradation pathway with the assumed activation energy; where the real activation energy differs materially, the calculated value misrepresents the exposure. It is meaningless for freezing damage, which is not an Arrhenius process. And it is a summary, so it cannot substitute for the profile when the question is whether a specific threshold was crossed.^[11][7]

Between controlled environments, the product travels inside a package that must maintain its internal temperature without external power. The performance of such a passive shipper is a designed property, qualified against defined ambient profiles, and it is specific to a payload.

The components are an insulating shell, a coolant, and a payload space arranged so that the coolant neither contacts the product directly nor is separated from it by so much air that it cannot influence it.

Insulation

Expanded polystyrene, polyurethane and vacuum-insulated panels in ascending order of thermal performance, cost and weight. Vacuum panels achieve roughly an order of magnitude lower conductivity than expanded polystyrene of the same thickness.

Coolant

Water ice, gel packs, phase-change materials formulated to melt at a chosen temperature, and dry ice for frozen shipments. Coolant mass and its conditioning before packing are as consequential as its identity.

Payload arrangement

A qualified configuration specifies coolant placement, payload volume and the minimum and maximum payload for which the qualification holds. A shipper qualified for a full payload frequently fails when shipped part-full, because the reduced thermal mass follows the coolant more closely.^[12]

Qualification is performed against standardised ambient profiles representing summer and winter conditions and including the transitions typical of air freight. ISTA Procedure 7D provides temperature test profiles for transport packaging, and ASTM D3103 gives a method for determining the thermal insulation quality of a package. Qualification establishes a duration — the time for which the package holds the payload within specification under the profile — and that duration is a property of the whole system, not of the insulation alone.^[13][14]

Water ice melts at 0 °C and therefore holds a payload close to that temperature while any ice remains — well below the 2 °C lower bound of the cold range, and a recognised cause of freezing excursions in 2–8 °C distribution. Phase-change materials are formulated to melt at a temperature inside the target range, commonly around 5 °C for refrigerated shipments and around 22 °C for controlled-room-temperature shipments, so that the latent heat of the transition buffers the payload at a temperature that is itself compliant.^[12]

The trade-off is that a material with a higher melting point stores its latent heat at a smaller temperature difference from ambient and therefore protects for a shorter time against a warm environment, while providing protection against freezing that water ice cannot. Where a shipment must survive both a hot and a cold ambient profile, dual-material configurations are used.^[4]

Failures cluster at a small number of predictable points, and the empirical literature is consistent about which.

Freezing. A 2007 systematic review of vaccine cold chains examined studies from high- and low-income settings and found evidence of freezing exposure in the majority of those examining shipments, and in a large proportion of those examining storage. Its conclusion — that inadvertent freezing was the more prevalent problem, and more damaging for the products concerned, than heat exposure — was surprising to practitioners at the time and has been sustained by a later review of the same question.^[4][5] The mechanisms are direct contact between payload and frozen coolant, refrigerators controlling to a set point without regard to compartment gradients, and domestic refrigerators used for storage, in which the coldest region may be several degrees below zero.

Custody transfers. Airport handling, customs holds and courier depots are the points at which a shipment is most likely to sit outside a controlled environment, and are also the points at which no party is monitoring. A hold of unpredictable duration at an unpredictable ambient temperature is the characteristic risk of cross-border movement, and it is the reason customs interventions have thermal consequences independent of their legal ones.^[9]

• Temperature excursion
• Cold-chain freight
• Peptide stability table
• Lyophilisation
• Beyond-use date
• Vial